• 0 Commentaires ·0 Parts ·50 Vue
  • Operative Games unveils AI-driven interactive storytelling platform
    venturebeat.com
    Operative Games, an interactive storytelling company, emerged today from stealth with backing from investors 1AM Gaming, Samsung Next, and LongJourney.vc.Read More
    0 Commentaires ·0 Parts ·57 Vue
  • ByteDance Research Releases DAPO: A Fully Open-Sourced LLM Reinforcement Learning System at Scale
    www.marktechpost.com
    Reinforcement learning (RL) has become central to advancing Large Language Models (LLMs), empowering them with improved reasoning capabilities necessary for complex tasks. However, the research community faces considerable challenges in reproducing state-of-the-art RL techniques due to incomplete disclosure of key training details by major industry players. This opacity has limited the progress of broader scientific efforts and collaborative research.Researchers from ByteDance, Tsinghua University, and the University of Hong Kong recently introduced DAPO (Dynamic Sampling Policy Optimization), an open-source large-scale reinforcement learning system designed for enhancing the reasoning abilities of Large Language Models. The DAPO system seeks to bridge the gap in reproducibility by openly sharing all algorithmic details, training procedures, and datasets. Built upon the verl framework, DAPO includes training codes and a thoroughly prepared dataset called DAPO-Math-17K, specifically designed for mathematical reasoning tasks.DAPOs technical foundation includes four core innovations aimed at resolving key challenges in reinforcement learning. The first, Clip-Higher, addresses the issue of entropy collapse, a situation where models prematurely settle into limited exploration patterns. By carefully managing the clipping ratio in policy updates, this technique encourages greater diversity in model outputs. Dynamic Sampling counters inefficiencies in training by dynamically filtering samples based on their usefulness, thus ensuring a more consistent gradient signal. The Token-level Policy Gradient Loss offers a refined loss calculation method, emphasizing token-level rather than sample-level adjustments to better accommodate varying lengths of reasoning sequences. Lastly, Overlong Reward Shaping introduces a controlled penalty for excessively long responses, gently guiding models toward concise and efficient reasoning.In practical experimentation, DAPO has demonstrated significant improvements. Evaluations on the American Invitational Mathematics Examination (AIME) 2024 benchmark show that DAPO-trained models achieved a score of 50 points using the Qwen2.5-32B base model, improving on previous methods such as DeepSeek-R1-Zero-Qwen-32B, which achieved 47 points. Notably, DAPO attained this improvement with approximately half the training steps, underscoring the efficiency of the proposed methods. A systematic analysis revealed incremental enhancements from each introduced technique, moving from a baseline of 30 points (using GRPO alone) up to 50 points with the full DAPO methodology.Beyond quantitative results, DAPOs training dynamics provided insights into the models evolving reasoning patterns. Initially, the models showed little reflective behavior, often proceeding linearly through tasks without reconsideration of previous steps. However, with ongoing training, the models progressively exhibited more reflective behaviors, demonstrating a form of iterative self-review. This shift highlights the capability of reinforcement learning not only to enhance existing reasoning pathways but also to cultivate entirely new cognitive strategies over time.In conclusion, the open-sourcing of DAPO represents a meaningful contribution to the reinforcement learning community, removing barriers previously created by inaccessible methodologies. By clearly documenting and providing comprehensive access to the systems techniques, dataset, and code, this collaborative initiative invites further research and innovation. The combined efforts of ByteDance, Tsinghua University, and the University of Hong Kong showcase the potential of transparent and cooperative research to advance the collective understanding and practical capabilities of large-scale reinforcement learning systems.Check outthe Paper and Project Page.All credit for this research goes to the researchers of this project. Also,feel free to follow us onTwitterand dont forget to join our80k+ ML SubReddit.The post ByteDance Research Releases DAPO: A Fully Open-Sourced LLM Reinforcement Learning System at Scale appeared first on MarkTechPost.
    0 Commentaires ·0 Parts ·57 Vue
  • Speech-to-Speech Foundation Models Pave the Way for Seamless Multilingual Interactions
    www.marktechpost.com
    At NVIDIA GTC25, Gnani.ai experts unveiled groundbreaking advancements in voice AI, focusing on the development and deployment of Speech-to-Speech Foundation Models. This innovative approach promises to overcome the limitations of traditional cascaded voice AI architectures, ushering in an era of seamless, multilingual, and emotionally aware voice interactions.The Limitations of Cascaded ArchitecturesCurrent state-of-the-art architecture powering voice agents involves a three-stage pipeline: Speech-to-Text (STT), Large Language Models (LLMs), and Text-to-Speech (TTS). While effective, this cascaded architecture suffers from significant drawbacks, primarily latency and error propagation. A cascaded architecture has multiple blocks in the pipeline, and each block will add its own latency. The cumulative latency across these stages can range from 2.5 to 3 seconds, leading to a poor user experience. Moreover, errors introduced in the STT stage propagate through the pipeline, compounding inaccuracies. This traditional architecture also loses critical paralinguistic features such as sentiment, emotion, and tone, resulting in monotonous and emotionally flat responses.Introducing Speech-to-Speech Foundation ModelsTo address these limitations, Gnani.ai presents a novel Speech-to-Speech Foundation Model. This model directly processes and generates audio, eliminating the need for intermediate text representations. The key innovation lies in training a massive audio encoder with 1.5 million hours of labeled data across 14 languages, capturing nuances of emotion, empathy, and tonality. This model employs a nested XL encoder, retrained with comprehensive data, and an input audio projector layer to map audio features into textual embeddings. For real-time streaming, audio and text features are interleaved, while non-streaming use cases utilize an embedding merge layer. The LLM layer, initially based on Llama 8B, was expanded to include 14 languages, necessitating the rebuilding of tokenizers. An output projector model generates mel spectrograms, enabling the creation of hyper-personalized voices.Key Benefits and Technical HurdlesThe Speech-to-Speech model offers several significant benefits. Firstly, it significantly reduces latency, moving from 2 seconds to approximately 850-900 milliseconds for the first token output. Secondly, it enhances accuracy by fusing ASR with the LLM layer, improving performance, especially for short and long speeches. Thirdly, the model achieves emotional awareness by capturing and modeling tonality, stress, and rate of speech. Fourthly, it enables improved interruption handling through contextual awareness, facilitating more natural interactions. Finally, the model is designed to handle low bandwidth audio effectively, which is crucial for telephony networks. Building this model presented several challenges, notably the massive data requirements. The team created a crowd-sourced system with 4 million users to generate emotionally rich conversational data. They also leveraged foundation models for synthetic data generation and trained on 13.5 million hours of publicly available data. The final model comprises a 9 billion parameter model, with 636 million for the audio input, 8 billion for the LLM, and 300 million for the TTS system.NVIDIAs Role in DevelopmentThe development of this model was heavily reliant on the NVIDIA stack. NVIDIA Nemo was used for training encoder-decoder models, and NeMo Curator facilitated synthetic text data generation. NVIDIA EVA was employed to generate audio pairs, combining proprietary information with synthetic data.Use CasesGnani.ai showcased two primary use cases: real-time language translation and customer support. The real-time language translation demo featured an AI engine facilitating a conversation between an English-speaking agent and a French-speaking customer. The customer support demo highlighted the models ability to handle cross-lingual conversations, interruptions, and emotional nuances.Speech-to-Speech Foundation ModelThe Speech-to-Speech Foundation Model represents a significant leap forward in voice AI. By eliminating the limitations of traditional architectures, this model enables more natural, efficient, and emotionally aware voice interactions. As the technology continues to evolve, it promises to transform various industries, from customer service to global communication. Jean-marc MommessinJean-marc is a successful AI business executive .He leads and accelerates growth for AI powered solutions and started a computer vision company in 2006. He is a recognized speaker at AI conferences and has an MBA from Stanford.Jean-marc Mommessinhttps://www.marktechpost.com/author/jean-marc0000677/Lowes Revolutionizes Retail with AI: From Personalized Shopping to Proactive Customer AssistanceJean-marc Mommessinhttps://www.marktechpost.com/author/jean-marc0000677/Google DeepMinds Gemini Robotics: Unleashing Embodied AI with Zero-Shot Control and Enhanced Spatial ReasoningJean-marc Mommessinhttps://www.marktechpost.com/author/jean-marc0000677/Aya Vision Unleashed: A Global AI Revolution in Multilingual Multimodal Power!Jean-marc Mommessinhttps://www.marktechpost.com/author/jean-marc0000677/Limbic AIs Generative AIEnabled Therapy Support Tool Improves Cognitive Behavioral Therapy Outcomes
    0 Commentaires ·0 Parts ·58 Vue
  • Bones and All: Ranking the Star Trek Doctors and Physicians
    www.denofgeek.com
    Given its mission to seek out new life and new civilizations, Star Trek doesnt give its physicians the most glamorous jobs. Captain Kirk and Commander Riker get to go on adventures, Picard and Spock wrestle with complex questions. The doctors just clean up the mess.And yet, the doctors remain some of the most intriguing characters in all of Trek. The advice they offer the captains, the moral dilemmas foisted upon them by the patients sent to their sick bays, even their romantic entanglements. All of these make for compelling stories, which might be why all of the Trek doctors are pretty great. Theres honestly not a bad one on this list, even if the shows around them are sometimes lacking.Still, there are a few caveats we need to put in place, which means some of the docs get left out. This list ranks only the main doctors, not other types of healers (sorry Nurses Chapel and Ogawa, sorry Counselor Troi) nor doctors who only show up for a bit. That means not only that were skipping over Dr. April from The Animated Series or Dr. Royce from The Cage, but also The Original Series version of MBenga played by Booker Bradshaw and the Kelvin universe McCoy played by Karl Urban. Finally, were just looking at the movies and TV shows here, so the friendly dinosaur Dr. Shenti Yisec Eres Ree from the Star Trek: Titan novels will not appear.Still, that leaves us with plenty of great physicians to look at, each of whom any Trekkie would be happy to visit.Photo: CBS Studios.9. Catherine Pulaski (TNG)Perhaps the easiest decision on this list, Dr. Pulaski takes the bottom spot. Yet, even that makes her sound worse than she is. Pulaski already has the odds stacked against her, coming in only for The Next Generations still-shaky second season and briefly replacing a fan-favorite in Beverly Crusher, when Diana Muldaur joined the cast after Gates McFaddens season one departure. Worse, she makes a terrible impression in her early appearances by bullying Data over the pronunciation of her name.However, the 24-episode seasons we used to have back then really worked in Pulaskis favor, giving her time to grow beyond those first impressions. By the time we get to the 15th episode, Pen Pals, Pulaski is showing remarkable compassion for a scared child, and even bonding with Data. Had Pulaski got to stick around through season three and longer, when TNG matured into the great show it would become, then perhaps she would be at the top of this list. As it is, she doesnt quite rank as high as the others.Photo: Paramount +.8. Tana (Lower Decks)The USS Cerritoss grouchy Caitian Dr. Tana underscores one of the key struggles of this list. When were considering what makes a great Star Trek doctor, are we giving more weight to their skills as a physician, and everything that entails? Or how much we enjoy watching them? Voiced with grit and vinegar by Gillian Vigman, TAna seems annoyed when anyone on the Ritos needs her help. Worse, she uses a very dodgy method of pain measurement. Basically, if it doesnt bother her and her incredibly high pain tolerance, then it should bother her patients either.That said, TAna is a delight to watch. Shes one of the more wacky side characters on Lower Decks and embodies the shows original intentions as a parody of Trek. Where Mariner and Boimler matured into full, real characters, TAna remains a send up of McCoy and Pulaski. Given her relatively little screen time, the enjoyment she brings to viewers isnt quite enough to outweigh the pain she brings to patients, and so TAna falls toward the bottom.7. Hugh Culber (Discovery)If Pulaski and TAna are a bit too rough on their patients, the Discoverys Dr. Hugh Culber falls way too far in the other direction. Actor Wilson Cruz has plenty of muscles and doesnt mind showing them off, but his greatest strength in Discovery is his kind smile. Culber is all empathy and understanding, which made him a good match for husband Paul Stamets, who started out as a pricklier character.However, as Discovery increasingly leaned into emotional storytelling, Culbers great empathy became less remarkable. In fact, after his death at the end of season one and resurrection midway through season two, Culber simply existed to give empathetic smiles to other characters, especially the dreadfully dull Adira, making him a compelling person who never realized his potential just like every other Discovery character. 6. EMH (Voyager)Like Dr. Tana, the Emergency Medical Hologram aka the Doctor from Voyager divides the audience and the crew. Because as much as we love watching Robert Picardo play the EMH, its hard not to feel bad for Voyagers crew. The EMH is endlessly arrogant, easily distracted by anyone who praises his horrible singing or holonovels. Yes, he knows how to treat his patients well, perhaps even better than biological physicians. But man, he seems exhausting.Join our mailing listGet the best of Den of Geek delivered right to your inbox!Seems is the operative word there. Because Picardo is a delight, making EMH one of the consistently best characters from episode one, a rarity on Voyager. When paired with Kes, Seven of Nine, Tom Paris (or even Andy Dick, in one episode), the EMH makes for a compelling foil. Furthermore, one need spend only a couple of minutes watching a medical show to see that arrogant, uncaring doctors can be great TV.5. Joseph MBenga (Strange New Worlds)Dr. Joseph MBenga sits right at the center of this list because he can go either direction. MBenga is an ideal character for Strange New Worlds, someone who showed up a couple of times in TOS, but never really had an established identity, giving the current writers plenty to work with. For the most part, writers have gotten it right with MBenga, taking full advantage of actor Babs Olusanmokuns distinctive voice and unique features to craft a person who exudes empathy and mystery.MBenga has been an MVP in several of SNWs best moments, whether hes excitedly recommending Captain Pikes breakfast offerings or reading a fairy tale to his daughter. Thus far, the show has handled his PTSD from the Klingon War well, making him a space age version of a Civil War physician seen in shows like Deadwood. However, the PTSD storyline also suggests that MBenga is actually a black ops super warrior, and even featured him becoming space Hulk for a few minutes. SNW has been mostly great so far, and MBenga has been a highlight. But if the supersoldier arc returns, its certainly going to knock him down a few pegs on this list.4. Julian Bashir (Deep Space Nine)With his soft features and gentle affect, Dr. Bashir seems a million miles away from the grizzled MBenga or frontier medic Bones. Which is ironic, since Bashir becomes the only doctor on this list who gets on-screen, in-canon experience as a war time physician. Of course, thats just one of several developments that occur across Deep Space Nines seven seasons, in which Bashir transforms from a cocky, womanizing wiz-kid to morally complex but fundamentally good-hearted professional.Its a testament to Alexander Siddig as an actor that he not only excelled at playing all these aspects of Bashir, but made them feel consistent. As Bashir learned about his augment status and developed a relationship with the secretive Garak, he also grew more caring, his pride knocked down a few pegs and his empathy growing. By the time the Dominion War reached its boiling point, Bashir established himself as the right doctor for the times.Photo: CBS Studios.3. Beverly Crusher (The Next Generation)In the same way that TNG turned its captain from the swashbuckling Kirk to the patrician Picard, so also did the series move away from McCoys gruff demeanor to the absolute professionalism of Dr. Beverly Crusher, played by Gates McFadden. In fact, Dr. Crusher exemplified TNGs emphasis on extreme competence better than any other main character, save perhaps Picard. Not even the fact that she was saddled with pre-Prodigy Wesley and a romantic plot with the Captain could distract from the fact that she is, ultimately, a doctor who is very, very good at her job.That professionalism meant that Crusher got to star in some of the best episodes of the series, courtesy not extended to any of her female co-stars (until Michelle Forbes joins as Ensign Ro). Remember Me and Ethics highlight Crushers moral compass and intelligence, and not even the oft-mocked Sub Rosa (which is better than its reputation suggests!) diminishes her as a person. Even better, when Crusher does finally return for season three of Star Trek: Picard, her romance storyline feels real and earned, as she has legitimate reasons for mistrusting Picard weighs them against her feelings. Crusher isnt just one of the best doctors in Star Trek, shes one of the best characters in the franchise.Photo: CBS Studios.2. Leonard Bones McCoy (TOS)Crusher might be more fully rounded and TAna might be funnier, but could anyone else on this list rock the beard, leisure suit, and medallion combo Bones wears in The Motion Picture? I think not. All joking aside, the incredible look and DeForest Kelleys mix of annoyance and respect when he realizes that Kirk has called him back into service is everything great about Dr. Leonard McCoy. Bones isnt the same professional as Crusher, nor is he necessarily the smartest or most decorated of the docs in the franchise, but he is one of the main reasons the franchise exists at all.Although he often gets overshadowed by his frienemy Spock, McCoy serves as the crucial third to the triad at the center of TOS. Where Spock represented logic and urged Kirk to think rationally, McCoy was all human passion, an interesting choice for a man of science. As a result, McCoy kept things grounded on the Enterprise. No matter how many aliens or space gods they would encounter, Bones would respond like any other human, with irritation and wonder and humor. Others on this list might be better at keeping people alive, but only Bones understands how to live.Photo: CBS Studios.1. Phlox (Enterprise)Honestly, Dr. Phlox deserves top spot just for not smothering Malcolm Reed the first chance he got. Throw in the fact that he had to serve under Captain Archer, and Phlox is an undisputed hero. In fact, Phlox rivals TPal and Trip Tucker as the only good main characters on Enterprise because he best captures the prequels concept, which followed humanitys first steps into deep space. Even more than TPal, Phlox served as Earths guide into deep space by modeling the mindset they need: deep curiosity and slow judgment.Played with even-keeled warmth by John Billingsley, Phlox meets every new challenge with wonder and kindness. He gently nudges his human shipmates toward new experiences (remember when he tried to get Trip to sleep with one of his wives), but never goes too far. Nor is he completely without morals, as hes more than willing to stand up against those who want to destroy other lifeforms. In short, Phlox isnt just a great Star Trek doctor. Hes a perfect Star Trek character, the one who best embodies the mandate to seek out new life and new civilizations.
    0 Commentaires ·0 Parts ·57 Vue
  • Microsoft Warns of StilachiRAT: A Stealthy RAT Targeting Credentials and Crypto Wallets
    thehackernews.com
    Mar 18, 2025Ravie LakshmananMalware / Threat IntelligenceMicrosoft is calling attention to a novel remote access trojan (RAT) named StilachiRAT that it said employs advanced techniques to sidestep detection and persist within target environments with an ultimate aim to steal sensitive data.The malware contains capabilities to "steal information from the target system, such as credentials stored in the browser, digital wallet information, data stored in the clipboard, as well as system information," the Microsoft Incident Response team said in an analysis.The tech giant said it discovered StilachiRAT in November 2024, with its RAT features present in a DLL module named "WWStartupCtrl64.dll." The malware has not been attributed to any specific threat actor or country.It's currently not clear how the malware is delivered to targets, but Microsoft noted that such trojans can be installed via various initial access routes, making it crucial for organizations to implement adequate security measures.StilachiRAT is designed to gather extensive system information, including operating system (OS) details, hardware identifiers like BIOS serial numbers, camera presence, active Remote Desktop Protocol (RDP) sessions, and running graphical user interface (GUI) applications.These details are collected through the Component Object Model (COM) Web-based Enterprise Management (WBEM) interfaces using WMI Query Language (WQL). It's also engineered to target a list of cryptocurrency wallet extensions installed within the Google Chrome web browser. The list encompasses Bitget Wallet, Trust Wallet, TronLink, MetaMask, TokenPocket, BNB Chain Wallet, OKX Wallet, Sui Wallet, Braavos Starknet Wallet, Coinbase Wallet, Leap Cosmos Wallet, Manta Wallet, Keplr, Phantom, Compass Wallet for Sei, Math Wallet, Fractal Wallet, Station Wallet, ConfluxPortal, and Plug.Furthermore, StilachiRAT extracts credentials stored in the Chrome browser, periodically collects clipboard content such as passwords and cryptocurrency wallets, monitors RDP sessions by capturing foreground window information, and establishes contact with a remote server to exfiltrate the harvested data.The command-and-control (C2) server communications are two-way, allowing the malware to launch instructions sent by it. The features point to a versatile tool for both espionage and system manipulation. As many as 10 different commands are supported -07 Display a dialog box with rendered HTML contents from a supplied URL08 - Clear event log entries09 - Enable system shutdown using an undocumented Windows API ("ntdll.dll!NtShutdownSystem")13 - Receive a network address from the C2 server and establish a new outbound connection.14 - Accept an incoming network connection on the supplied TCP port15 - Terminate open network connections16 - Launch a specified application19 - Enumerate all open windows of the current desktop to search for a requested title bar text26 - Put the system into either a suspended (sleep) state or hibernation30 - Steal Google Chrome passwords"StilachiRAT displays anti-forensic behavior by clearing event logs and checking certain system conditions to evade detection," Microsoft said. "This includes looping checks for analysis tools and sandbox timers that prevent its full activation in virtual environments commonly used for malware analysis."The disclosure comes as Palo Alto Networks Unit 42 detailed three unusual malware samples that it detected last year, counting a passive Internet Information Services (IIS) backdoor developed in C++/CLI, a bootkit that uses an unsecured kernel driver to install a GRUB 2 bootloader, and a Windows implant of a cross-platform post-exploitation framework developed in C++ called ProjectGeass.The IIS backdoor is equipped to parse certain incoming HTTP requests containing a predefined header and execute the commands within them, granting it the ability to run commands, get system metadata, create new processes, execute PowerShell code, and inject shellcode into a running or new process. The bootkit, on the other hand, is a 64-bit DLL that installs a GRUB 2 bootloader disk image by means of a legitimately signed kernel driver named ampa.sys. It's assessed to be a proof-of-concept (PoC) created by unknown parties from the University of Mississippi."When rebooted, the GRUB 2 bootloader shows an image and periodically plays Dixie through the PC speaker. This behavior could indicate that the malware is an offensive prank," Unit 42 researcher Dominik Reichel said. "Notably, patching a system with this customized GRUB 2 bootloader image of the malware only works on certain disk configurations."Found this article interesting? Follow us on Twitter and LinkedIn to read more exclusive content we post.SHARE
    0 Commentaires ·0 Parts ·58 Vue
  • Spiral Roll Geometry Node Presets - Motion Presets Update [$]
    www.blendernation.com
    Spiral Roll Geometry Node Presets - Motion Presets Update By Blender Renaissance on March 18, 2025 Add-ons The March 2025 update for Motion Presets for Blender has been released, here are the new features.Spiral Node Geometry Node Presets: You can roll a plane mesh either horizontally or vertically Node Setup is programmatically generated by the python script, no extra blend files are needed 1 Click option for Mesh preparation that gives you greater control of the mesh Un-realized subdivisions, helps for efficiency in memory and performance Animate Multiple meshes at once, each mesh has its own separate Geometry Node and Animation data 16 Geometry Node Presets for both standing meshes and flat meshes, no need to worry about axes of the meshes Can be used with existing Motion Presets Thanking you,Vikrant JadhavBlender RenaissanceLinks
    0 Commentaires ·0 Parts ·59 Vue
  • The Art of Geoffroy Thoorens
    www.iamag.co
    cookielawinfo-checkbox-analytics11 monthsThis cookie is set by GDPR Cookie Consent plugin. The cookie is used to store the user consent for the cookies in the category "Analytics".cookielawinfo-checkbox-functional11 monthsThe cookie is set by GDPR cookie consent to record the user consent for the cookies in the category "Functional".cookielawinfo-checkbox-necessary11 monthsThis cookie is set by GDPR Cookie Consent plugin. The cookies is used to store the user consent for the cookies in the category "Necessary".cookielawinfo-checkbox-others11 monthsThis cookie is set by GDPR Cookie Consent plugin. The cookie is used to store the user consent for the cookies in the category "Other.cookielawinfo-checkbox-performance11 monthsThis cookie is set by GDPR Cookie Consent plugin. The cookie is used to store the user consent for the cookies in the category "Performance".viewed_cookie_policy11 monthsThe cookie is set by the GDPR Cookie Consent plugin and is used to store whether or not user has consented to the use of cookies. It does not store any personal data.
    0 Commentaires ·0 Parts ·73 Vue
  • Google, Speedinvest back Kenyas Leta, which uses AI to make logistics cheaper
    techcrunch.com
    African businesses pay up to four times the global average to transport goods, driving up prices for essentials like food and medicine. Logistics make up 75% of product costs on the continent, according to the African Development Bank (AfDB). Many of these businesses also rely on manual logistics, leading to delays.Nairobi-based logistics software-as-a-service provider Leta wants to change all of that. Its AI-powered platform optimizes delivery routes, tracks shipments in real time, streamlines payments, and provides businesses with shipping insights.The startup has raised $5 million in seed funding to scale its solution, which it says is helping businesses move goods cheaper and faster across the continent. European VC firm Speedinvest led the round, with backing from Googles Africa Investment Fund and Equator, an Africa-focused climate tech fund.In November 2022, the Kenyan logistics startup raised a $3 million pre-seed from several local investors, money it used to deepen operations in its five core markets: Kenya, Nigeria, Uganda, Zambia and Zimbabwe. Leta, a Kenyan supply chain and logistics SaaS provider, raises $3M to scale in AfricaLetas load and route optimization technology helps its clients cut costs and improve delivery efficiency by reducing the number of vehicles needed for distribution, Founder and CEO Nick Joshi explains.Leta integrates directly with businesses ERP, POS, and OMS systems, pulling in live order data like SKUs, product types, prices, and customer details, Joshi says.From there, the platform selects the best available vehicle for each order and decides whether to load products using first-in, first-out (FIFO) or last-in, first-out (LIFO) methods, replacing manual, intuition-based dispatching. (FIFO loads the oldest inventory first, while LIFO loads the most recent stock first.)The platform then automates manifest creation and dispatch planning, optimizing vehicle use based on regional demand and truck capacity. Finally, Letas system, which, according to Joshi, is powered by AI, optimizes delivery routes in real-time.For example, if theres a roundabout where trucks or motorbikes repeatedly fail to complete a turn on that route, the AI flags it as a blacklisted route, said the CEO. It could be due to flooding, police stops, construction, or a presidential convoy. The system constantly updates its map layer to reflect these changes.Logistics, embedded finance, and sustainability playsLetas real-time mapping has become a key asset for Google, one of its investors. Joshi notes that Google Maps hasnt updated some areas of Nairobi since 2022, whereas Letas platform continuously refines road and address data sourced from live customer deliveries.Were creating a much more robust map and address layout, which is why I think Google found it interesting, he explains. By connecting stakeholders across the supply chain, Joshi sees financial services as a natural extension of Letas software platform and is already piloting some new products. Joshi says potential offerings include fuel cards for delivery partners, asset financing for vehicles and devices, and supply chain financing for FMCG merchants.Deepali Nangia, who leads Speedinvests investments in Africa and the Middle East, said the firm backed Leta because it leverages logistics as a gateway and fintech as a growth driver, unlocking new business opportunities.Equator closes $55M fund to bring more private capital to African climate techLeta also helps businesses reduce fleet sizes without cutting deliveries, lowering fuel consumption and emissions, which explains Equators backing. A company with 70 trucks saves about $30,000 monthly using Leta, Joshi claims. We havent started tracking carbon emissions yet, but its a key goal for this year.The Kenyan startup now powers 35+ major businesses, including global brands like KFC and Diageo, and local giants like EABL and Gilani, optimizing 10,000+ daily trips across its five markets.Since our 2022 coverage, Leta has seen massive growth: 500,000 deliveries to 4.5 million, from moving 20,000 tons to 150,000, and from managing 2,000 vehicles to 7,400. As a result, Letas revenues, which it makes on a per-delivery pricing model, have grown 5x, says Joshi.Leta now aims to double revenue in the coming months as it expands into more countries across Africa and the Middle East with clients like KFC and Diageo.Globally, Leta mirrors early Flexport before it shifted into tech-enabled fulfillment and asset ownership. In Africa, logistics startups like Sendy, Lori (also backed by Google) and KOBO360 took an asset-heavy approach, aggregating trucks and acting as intermediaries. However, this model has struggled, leading to recent closures and pivots.Leta takes a different approach: just software. Instead of owning or aggregating assets, it partners with companies that already own fleets, helping them boost efficiency and optimize utilization. Its a playbook other global logistics tech firms like Bringg, Onfleet, and Shipsy also follow.The first generation of logistics startups in Africa did the hard work by educating the market and proving whats possible, says Joshi. By the time we entered, some were exiting or trying to redefine their business. So we knew then what the market was looking for and what they needed.
    0 Commentaires ·0 Parts ·58 Vue
  • New 3D Printing Jobs at Ai Build, Norco, Velo3D, WAAM3D, and more
    3dprintingindustry.com
    Welcome to the latest edition of our 3D printing jobs and career moves update for the additive manufacturing sector.In this edition, well highlight recent developments and movements in the industrys workforce, shedding light on the dynamic landscape of the additive manufacturing sector.Read on for recent hires and facility openings at Ai Build, Materialise, Norco, Nikon Corporation, WAAM3D and more.New Hires at Norco, SWISSto12, Velo3D, WAAM3D, and moreKicking off with new hires, large composites, and GRP moldings manufacturer Norco has welcomed Max Osmond as the new head of Additive Manufacturing and CNC Development, reinforcing efforts to expand in advanced manufacturing. With a 27-year tenure as Director of MSA Manufacturing Limited, Osmond brings deep expertise in CNC kitting, 3D printing, and composite machining.Based in Ferndown, Dorset, MSA Manufacturing played a vital role in supplying composite materials before ceasing operations in 2024. At Norco, Osmond will focus on strengthening large-format 3D printing, streamlining CNC and additive workflows, and scaling operations. Industry relationships and technical knowledge position him to drive efficiency and innovation, with a long-term vision centered on team development and refining manufacturing processes for sustained growth.Max Osmond. Photo via Norco.Next up, Ben Morgan has stepped into the role of interim CEO at the Advanced Manufacturing Research Centre (AMRC), following the departure of Steve Foxley in December 2024.Previously serving as AMRCs research director and chief technology officer (CTO), Morgan will now lead the AMRC Board, collaborating with industrial partners, government bodies, and the wider team across South Yorkshire, Wales, and Lancashire. Since joining AMRC as a project engineer in 2009, Morgan progressed through key leadership roles, including head of the integrated manufacturing group at Factory 2050, where he led a team of over 70, and later as research director for more than five years.Ben said, I am excited and humbled to lead a wonderful organisation for which I have devoted my career. I look forward to working with the AMRC team, the wider University and colleagues across the High Value Manufacturing Catapult to continue to deliver impact to industry across the UK.Cranfield University spin-out WAAM3D has appointed Lee Chee Weng as its interim CEO at WAAM3D, following Filomeno Martinas decision to step down after six years of leadership. As a long-time board member, Lee also serves as Managing Director of Addept3D and Vice President of Advanced Manufacturing at Accuron.Martinas journey with WAAM3D began in 2010 as a Cranfield University MSc student, driven by a belief in the potential of wire arc additive manufacturing. Over the years, he helped grow the company from an academic team into a commercial enterprise, navigating challenges and expanding operations to Milton Keynes. While stepping away from day-to-day leadership, he remains a shareholder and looks forward to WAAM3Ds continued success under Lees guidance. Filomeno Martina, WAAM3D CEO, with the new MiniWAAM 3D printer at TCT3Sixty 2024. Photo by 3D Printing Industry.Swiss technology group Oerlikon has nominated Dr. Stefan Brupbacher, Marco Musetti, and Dr. Eveline Steinberger for election to its Board of Directors at the upcoming Annual General Meeting (AGM) on April 1, 2025. If approved, they will replace Irina Matveeva, Gerhard Pegam, and Zhenguo Yao, who are stepping down.With a strong background in economic policy and industry leadership, Brupbacher brings experience as Director of Swissmem and a board member at Orgalim. His previous role as Secretary General of the Swiss Federal Department of Economic Affairs, Education, and Research (EAER) involved shaping policies that supported Switzerlands industrial and technological growth. Plans are in place for him to take on the role of Lead Independent Director and chair the Governance Committee.Moreover, Musetti has spent years working across industries, holding board positions at medmix AG, Octo Telematics, GEM Capital, and United Kalahari Minerals. His expertise in corporate governance, business development, and investment strategy makes him a strong candidate for the Audit & Finance Committee. Additionally, his international experience adds a global perspective to the boards decision-making.As a specialist in AI, big data, augmented reality, and robotics, Steinberger founded The Blue Minds Company in 2014, focusing on digitalization and energy transition. Leadership roles at Siemens, the Climate and Energy Fund of the Austrian Federal Government, and VERBUND AG highlight her background in energy and infrastructure. As a Supervisory Board member at UniCredit Bank Austria, she brings additional financial and strategic expertise, with plans for her to serve on the Governance and Human Resources Committees.Apart from these changes, Oerlikons board structure remains unchanged, ensuring continuity while integrating fresh perspectives into its leadership team.Radio Frequency (RF) component supplier SWISSto12 has named Steve Collar as Chairman of its Board of Directors, bringing more than 30 years of experience in the satellite industry. His leadership at satellite communication service firm SES included five years as CEO, where he oversaw one of the worlds largest satellite operators.Before that, he spent over six years at O3b Networks, guiding the company from its early funding stages to a successful satellite constellation launch and commercial expansion. Under his direction, O3b became the fastest-growing satellite operator in the industry. Stepping into the role previously held by Roland Loos, who will remain on the board, Collar is set to contribute his expertise in scaling satellite businesses and driving industry growth.America Makes has appointed Andrew Thompson as chair of the Roadmap Advisory Group (RMAG) and Rick Russell as chairperson of the Executive Committee, bringing experienced leadership to guide the institutes strategic initiatives.Serving as Deputy Chief Engineer for Additive Manufacturing at Northrop Grumman Space Systems, Thompson will take charge of shaping RMAGs direction. His role involves leading key initiatives, recruiting members, engaging working groups, and ensuring the roadmap stays aligned with America Makes mission. He will also collaborate with the institute to maintain data, set priorities, and keep the Executive Committee informed on project call topics.As a chairperson, Russell will oversee the Executive Committee, working alongside industry experts from academia, government, and economic development sectors. His leadership will focus on implementing strategies, policies, and advocacy efforts that accelerate additive manufacturing adoption and enhance U.S. manufacturing competitiveness on a global scale.Supporting Russells efforts, Sandra DeVincent Wolf, Ph.D., has been appointed Secretary of the Executive Committee. As Executive Director of Carnegie Mellon Universitys (CMU) Manufacturing Futures Institute, she plays a key role in advancing manufacturing research. She also leads the NextManufacturing Center, a major hub for additive manufacturing research, where she manages partnerships, industry consortiums, and metals AM laboratories.Software development company ModuleWorks has appointed Amod Onkar as the Global Head of Marketing. He previously served as the Country Manager of India for SolidCAM. With over 25 years of experience in CAD/CAM and CNC machining, he brings expertise in strategic business development, team leadership, and industry innovation.During his tenure at SolidCAM, Amod expanded the companys customer base from two to over 1,200, driving a 20% annual growth rate in the past five years. He led a team of 72 engineers and sales professionals while advancing 3- and 5-axis milling technologies. His experience integrating ModuleWorks toolpath engines into SolidCAM products provides valuable insight for overseeing ModuleWorks global marketing and expansion efforts.I have known Amod for the past 20 years and could appreciate his work the whole time, especially his deep knowledge of CAM, says Yavuz Murtezaoglu, Managing Director and founder of ModuleWorks.Amod Onkar at ModuleWorks India. Photo via ModuleWorks.The Photopolymer Additive Manufacturing Alliance (PAMA) has appointed Vince Anewenter, Director of the Rapid Prototyping Consortium at the Milwaukee School of Engineering (MSOE), as the new Chair of its Executive Advisory Board. He takes over from David Walker, CTO of PrintFoam and Co-Founder of Azul 3D, who will continue with PAMA as Executive Advisory Board Chair, Ex-Officio.PAMA has also announced that Dr. Callie Higgins, a leading researcher at the National Institute of Standards and Technology (NIST), will step into the role in January 2027. She currently serves as the AM Coordinator for NISTs Material Measurement Laboratory and co-leads the Photopolymer Additive Manufacturing Project.Functioning as a collaboration initiative between NIST and RadTech, PAMA is dedicated to establishing industry-wide standards for photopolymer additive manufacturing. In 2025, the organization plans to focus on standardizing industry terminology and addressing the growing use of photopolymers in consumer and maker markets, particularly in terms of safety, health, and environmental considerations.Metal 3D printer manufacturer Velo3D has appointed Darren Beckett as its new CTO, entrusting him with guiding the companys technological direction and innovation efforts.In his new position at Velo3D, Beckett will oversee the development and execution of technology initiatives, ensuring they support the companys growth and align with industry needs. Having spent nearly 20 years at Intel Corporation, Beckett brings over 25 years of experience in technology leadership.His career also includes leadership roles at Woodruff Scientific, where he served as VP of Engineering, and Sigma Additive Solutions, where he worked as CTO. At Sigma, he played a key role in advancing additive manufacturing quality control technologies.New facility openings by AML3D, Materialise, Renishaw, Nikon, and moreMoving on to facility openings, Australian metal 3D printer manufacturer AML3D has launched its fully operational US technology center in Stow, Ohio, serving as its headquarters and a manufacturing hub.The facility currently houses an ARCEMY AM system for large-scale metal 3D printing, with a second system set for installation in early 2025. Stow is fulfilling a AUD$2.27 million order from the Tennessee Valley Authority (TVA), while AML3D plans to invest AUD$12 million to expand capacity in response to rising U.S. Department of Defense (DoD) demand, particularly for the US Navys Submarine Industrial Base.Led by Pete Goumas, President of US Operations, Stow also supports AML3Ds sales, administration, and technical teams. The expansion strengthens AML3Ds US Scale up strategy, which has secured over AUD$16 million in contracts since 2023, and enhances its ability to access ITAR-restricted defense contracts while positioning the company for potential US policy shifts favoring domestic manufacturing.Oil & gas company Aker Solutions has launched a 3D printing center in Trondheim to explore the potential of additive manufacturing (AM) for offshore maintenance and modifications. Developed together with Equinor, the ADDMO center is part of a seven-month pilot contract supporting operations at the Troll and Heidrun fields.At the center, engineers will work on integrating AM into daily workflows, focusing on reducing costs, improving efficiency, and lowering emissions by enabling on-demand production of spare parts. This initiative builds on experience from the Johan Castberg project, where 7,000 components were successfully 3D printed. With more than 50 AM systems now in use across multiple locations, Aker Solutions continues to expand the role of 3D printing in offshore asset management.Aker Solutions CEO Kjetel Digre (left) and EVP Life Cycle Paal Eikeseth at the opening of Aker Solutions 3D printing center ADDMO in Trondheim. Photo via Aker Solutions.London-based 3D printing software developer Aibuild has opened a new office in Silicon Valley to support its AI-powered AM operations in the U.S. Located within NIKONs research campus in the San Francisco Bay Area, the office will facilitate alliance between the two companies, with NIKON being one of Aibuilds key investors.This new site will serve as a research and development hub, allowing Aibuild to expand its customer base and form new partnerships with U.S. manufacturers. The company aims to contribute to supply chain sovereignty in industries such as aerospace, defense, and energy, which are considered critical to national security.Daghan Cam and Michail Desyllas, the founders of Aibuild said, Expanding into United States with a new office in Silicon Valley is a pivotal moment for Aibuild. Getting closer to our key customers and partners in the region will allow us to respond to their needs faster. Working alongside NIKON in our new location will also deepen our partnership and accelerate our efforts to bring AI-driven automation into manufacturing.Additionally, Nikon Corporation is set to open the Nikon AM Technology Center Japan in Gyoda, Saitama Prefecture, with a grand opening event planned for February 28, 2025. Following the launch of its AM Technology Center in Long Beach, California, in July 2024, the new facility will serve as a hub for advanced manufacturing in Japan and Asia.Equipped with Nikon SLM Solutions NXG XII 600 laser powder bed fusion (PBF) system, the first of its kind in Japan, the center will focus on metal additive manufacturing for defense, space, and aviation sectors. High-precision directed energy deposition (DED) systems, including the Lasermeister LM300A, will also be available for repair and sustainment applications across various industries.Global engineering firm Renishaw and IDEKO have opened a new Renishaw Solutions Centre in Elgoibar, Spain, to support research in advanced manufacturing. Located within the Basque Research and Technology Alliance (BRTA), the facility is part of a partnership agreement signed during the 2024 International Machine Tool Exhibition in Bilbao.Industry leaders, including Deputy Minister for Technology, Innovation, and Digital Transformation Jaione Ganzarain, attended the inauguration, where technologies for 3D measurement, process control, metal AM, and industrial robotics calibration were presented. Additionally, the partnership extends to machine tools manufacturer Danobatgroup alongside the engineering firm supplying advanced technologies to enhance manufacturing capabilities across multiple product lines.Renishaw and IDEKO team cutting the ribbon. Photo via Renishaw.Elsewhere, Belgian 3D printing software and services provider Materialise has opened an Aerospace Competence Center in the Aerospace Innovation Hub in Delft, becoming the first additive manufacturing (AM) company to establish a presence there. The company has produced over 500,000 3D printed aerospace parts annually, supporting OEMs, airlines, MROs, and supplier tiers.Establishing a base in Delft allows Materialise to work with TU Delft, startups, and industry professionals, contributing to research in sustainable aviation and advanced manufacturing. With EASA POA and EN 9100 certifications, the company is positioned to support new AM applications while improving supply chain efficiency, lead times, and part design flexibility. Industrial giants Airbus and Collins Aerospace are already part of the innovation hub, fostering broader partnerships in aerospace technology.The opening of our new Aerospace Competence Center aligns with our commitment to support the aerospace industry through more than three decades of experience in additive manufacturing and software solutions, as well as our pioneering role in producing certified parts, said Brigitte de Vet-Veithen, CEO of Materialise.What3D printing trendsshould you watch out for in 2025?How is thefuture of 3D printingshaping up?To stay up to date with the latest 3D printing news, dont forget to subscribe to the 3D Printing Industry newsletter or follow us on Twitter, or like our page on Facebook.While youre here, why not subscribe to our Youtube channel? Featuring discussion, debriefs, video shorts, and webinar replays.Featured image shows 3D Printing Industry Jobs Board. Image via 3D Printing Industry.Ada ShaikhnagWith a background in journalism, Ada has a keen interest in frontier technology and its application in the wider world. Ada reports on aspects of 3D printing ranging from aerospace and automotive to medical and dental.
    0 Commentaires ·0 Parts ·60 Vue