Sign in to view Reza’s full profile
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Sign in to view Reza’s full profile
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Stanford, California, United States
Sign in to view Reza’s full profile
Reza can introduce you to 10+ people at Matroid
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
9K followers
500+ connections
Sign in to view Reza’s full profile
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
View mutual connections with Reza
Reza can introduce you to 10+ people at Matroid
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
View mutual connections with Reza
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Sign in to view Reza’s full profile
or
New to LinkedIn? Join now
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
About
Welcome back
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
New to LinkedIn? Join now
Activity
9K followers
-
Reza Zadeh reposted thisReza Zadeh reposted thisMatroid's CEO, Reza Zadeh, met again with His Majesty King Charles III at the Sustainable Markets Initiative Terra Carta Roundtable in March 2026. The conversations continue to move beyond vision and into execution. Artificial intelligence is now part of how industries operate, how infrastructure is managed, and how sustainability goals are achieved. At Matroid, we are seeing this shift every day across manufacturing, energy, and aerospace. Computer vision is turning visual data into real-time decisions that improve quality, safety, and efficiency at scale. It is meaningful to see this work aligned with global leadership focused on long-term impact. Grateful for the continued dialogue and leadership. #TerraCarta #SMI #AI #Sustainability
-
Reza Zadeh reposted thisReza Zadeh reposted thisInnovation is a mirror of our priorities. That enduring truth echoed in the halls of Hampton Court Palace in London this week. This year’s Sustainable Markets Initiative Roundtables & Exhibition celebrated the 175th anniversary of another exhibition, the 1851 "Great Exhibition." That landmark event of the 19th century brought together more than 100,000 “Works of Industry of All Nations” inside the famed Crystal Palace. It highlighted major advances in engineering and manufacturing and the industrial capacity of the Victorian era. It also reflected the belief of Prince Albert, the fair's royal champion, that technology could support broader social and economic aims. Albert’s vision still resonates. Moderating the SMI’s AI for Transition plenary yesterday gave me the chance to speak with the CEOs of AECOM, ExpectAI, Matroid and Regrow Ag. Their work shows how AI is helping to reduce energy demands, improve design, strengthen decision‑making in complex operations and enable more sustainable land use at scale. Their experiences underscored the importance of high-quality data, contextual understanding and responsible governance as AI becomes embedded in industry. As a “coalition of the willing,” the SMI can showcase progress, democratize adoption and accelerate “net-positive” AI impacts. Unlike 1851, our innovation priority today is not to build bigger and more powerful machines to master nature and the physical world. Rather, it is to harness data-driven insights—enriched with AI—to make energy, water, infrastructure and agricultural systems more resilient and efficient. #TheSMI #IndustrialIntelligence #SustainableTransition Anand Verma (Dr.) Anastasia Volkova, PhD Reza Zadeh Troy Rudd Sophie Miremadi Jacob Alexander Megan Lehtonen
-
Reza Zadeh reposted thisReza Zadeh reposted thisAt Scaled ML, one idea from Ashok Elluswamy stood out. “AI is the driver.” At Tesla, AI is literally the driver. It powers the foundational models that see the world, make decisions, and operate vehicles in real time. That same idea extends beyond autonomy. AI is becoming a driver of how humans build systems, solve problems, and extend our own capabilities. It is shaping how we move through the world and how progress compounds across generations. In this clip from Ashok’s Scaled ML talk, titled “Building Foundational Models for Robotics,” he explains how Tesla approaches foundational models and why that perspective matters for what comes next. Watch the full talk on YouTube here: https://lnkd.in/gmJGSiip #ScaledML #AI #Autonomy #FoundationalModels #MachineLearning
-
Reza Zadeh reposted thisReza Zadeh reposted thisDatabricks co-founders Matei Zaharia and Ion Stoica will be speaking at #ScaledML 2026! This event brings together the creators behind systems like Apache Spark™, CUDA, TensorFlow, ImageNet, Tesla FSD, and more to explore large-scale learning, distributed systems, and next-generation AI hardware. Join us on January 29th: https://scaledml.org/
-
Reza Zadeh posted thisThe alpha in ScaledML companies is insane: Groq, Cerebras, OpenAI, Tesla, Google, Meta, Databricks, Matroid and many others have all at least 10x'ed in valuation since first presenting at ScaledML. ScaledML.org
-
Reza Zadeh shared thisScaledML is back! After a 6 year hiatus. Consistently 2-3 years ahead of where ML will be. Examples of foresight at SML: - OpenAI in 2016,2017,2019 announced GPT-2 & RL efforts - Turing award for Deep Learning announced by Turing award winner on morning of award - Groq chip (acquired for $20bn) released - All Google TPU versions detailed Join us on January 29th at the Computer History Museum! http://scaledml.org
-
Reza Zadeh reposted thisExcited to present at #ScaledML 2026. I'll share perspectives on what we've been working on at the DAF-Stanford AI Studio to scale AI/ML to solve some of the toughest technical problems for the United States Air Force and United States Space Force.Reza Zadeh reposted thisScaledML is back! The world’s leading forum for large-scale machine learning, distributed systems, and real-world AI engineering returns with an exceptional lineup of researchers, founders, and industry practitioners. Seats are limited. Secure yours today with discount code 2026SMLEarlyBird https://lnkd.in/gxUjgZuQ #ScaledML #AI #ML #CuttingEdge #Conference
-
Reza Zadeh reposted thisReza Zadeh reposted thisWe disclosed today as part of our Series L that our 4-yr old data warehousing business is now >$1B revenue run rate. This is to the best of my knowledge the fastest to $1B DW product in the industry. The conventional wisdom is that it would take 5+ years to build a new database (just to release one). How did we do it, and what’s next? Four years ago, the linked blog announced that Databricks had won the official TPC-DS 100TB benchmark with DBSQL, which was in preview back then. It had the best perf and the best price/perf, and notably beating Snowflake by 12x in price/perf in that benchmark. (Note: we are still the top place on the official TPC-DS benchmark today.) That blog post launched a contentious "benchmarking war" with a lot of back and forth between vendors, but more importantly it marked the very beginning of our data warehousing business. To build this business, we assembled the best engineering team and established a new infrastructure product category called Lakehouse that inherits the flexibility and openness of data lakes and performance of data warehouses. Lakehouse is now the standard for data infrastructure, and organizations are migrating from legacy data warehouses to the Lakehouse. The result so far is a testament to the team and their execution. We have a lot of ideas on how to take performance and usability to the next level, and the team is working hard to make that happen. Expect some big announcements next year. We want to lay the foundation for growing the data warehousing product to a $10B business. Databricks had operated largely in the “analytics” side of data in the past, and we believe the “operational” side of data (aka “OLTP”) is also ready for a “Lakehouse” style disruption. A huge chunk of the founding team’s time is now focusing on “Lakebase”, a new category of OLTP databases that separates storage (in the lake) from compute. That architecture enables features that have been virtually impossible for databases in the past: instant provisioning, elastic scaling (down to zero), branching, high throughput scan directly from Spark, … I won’t go into too much detail about Lakebase here, but we expect a similar trend to happen in the next few years: Lakebase will transform the industry and other OLTP systems will re-architect or position towards it. The best data warehouse is a lakehouse, and the best database is a lakebase! https://lnkd.in/gjhzMfAtDatabricks Sets Official Data Warehousing Performance RecordDatabricks Sets Official Data Warehousing Performance Record
-
Reza Zadeh reposted thisReza Zadeh reposted thisScaledML is back! Over the years, ScaledML has always featured incredible talks. As the official event photographer, I've witnessed some pretty cool moments: - Ilya Sutskever shared outputs from early GPT models in 2018 (turns out he was onto something). - Jeff Dean introduced a new framework called TensorFlow and discussed how Google was beginning to use ML more broadly back in 2016 - In 2020 Andrej Karpathy dove into the nuances of stop-sign recognition -- for example, what should a model do if a construction worker is holding a portable stop sign, but their arm is down and the sign is upside-down by their hip? - Dave Patterson announced the 2019 Turing Award for Yoshua Bengio, Geoff Hinton, and Yann LeCun -- a milestone moment for deep learning ScaledML has always been a glimpse into the future -- consistently a few steps ahead of where the industry is heading. If you're interested in joining us this year, registration is now open: https://scaledml.org
-
Reza Zadeh liked thisReza Zadeh liked this❤️❤️Today, Databricks is proud to announce an >$850m investment in the #UK🇬🇧over the next few years. 🚀🚀 This will support the expansion into a brand new #EMEA hub office in central #London, double our team, expand our #R&D efforts, accelerate #AI #training and support our customers to innovate with #Lakebase and #Genie. The UK has one of the most exciting AI ecosystems in #Europe and we are proud to be part of its growth. Congratulations to our whole UKI team, Michael Green and thank you to our partners and customers for your trust. Check out the full story 👉https://lnkd.in/dV993Exx
-
Reza Zadeh liked thisReza Zadeh liked thisThe ACT4Aero agenda is shaping up nicely! Thrilled to announce some of our panels and firesides on Day 1, defining opportunities and thresholds of success - Harnessing Emerging AI Technologies for National Competitiveness w/ Patrick Lincoln (DARPA) and Charles Ball (Genesis Mission). - Closing the Loop between S&T and T&E w/ Maj Gen Scott Cain (AF Test Center) and Victoria Coleman (Berkeley) - Modernizing Test and Acquisitions for Physical AI Age w/ Lt Col Raven "Rost" Leclair and Hon Dale Marks To see the rest of the agenda, visit https://act4aero.com. Huge thanks to all of our panelists and special thanks to the rest of the organizing committee: Jason Hansberger, Charbel Farhat, Marco Pavone, Juan Alonso, and Chris Dylewski. Excited to see all of our collaborators at the event.
-
Reza Zadeh reacted on thisReza Zadeh reacted on thisLast night in London was a powerful reminder of what committed people can achieve. Honoured to attend the TIME Earth Awards 🌍, bringing together inspiring leaders and changemakers committed to driving change. A special moment seeing SMI Ambassador Stella McCartney recognised for her continued leadership in sustainable innovation. Her work continues to show how business, creativity, and responsibility can intersect to drive real impact. Grateful to have been in the room with so many individuals pushing the agenda forward. Events like this reinforce that progress happens when we collaborate, challenge norms, and stay committed to change to sieze #thegrowthstoryofourtime. Sustainable Markets Initiative #Sustainability #Leadership #SMI #TimeEarthAwards #SMITerraCarta
-
Reza Zadeh liked thisReza Zadeh liked this"An astonishing display of visual storytelling." (Harvard Business Review) Astronaut. Test pilot. Commander. Bestselling author. Speaking globally on leadership, high-stakes decision-making, risk, and performing at your best when it matters most. #KeynoteSpeaker #GlobalSpeaker #Astronaut
-
Reza Zadeh liked thisReza Zadeh liked thisThis time, I believe it truly is the last time. This is my last week at Meta, and I couldn't be more excited about what's next. I'll keep this short as, for those who know me, I've moved away from getting sentimental on public channels and prefer connecting with colleagues I've worked closely with in a more personal way. A few parting thoughts I had for my colleagues inside Meta - coming from a genuine place: 1. The people and relationships are what matter most - The company will always survive, but we spend more time with each other than with our actual families. While we do what's right for those who sign our paychecks, the relationships we build and friendships we make are what truly matter. 2. Align your energy and passion with what you do - Maybe I'm just getting older and less concerned with compensation, vesting schedules, etc., but I've made decisions on roles, projects, and companies for the wrong reasons. If you're excited about something, do your best to make it your job. A good litmus test: what do you spend your personal time on - nights and weekends? If work becomes a chore, it's time to move on. 3. AI is changing everything, BUT - There's definitely good happening with companies and people building solutions with impact beyond driving ad revenue or social media engagement. Think drug discovery, finding new materials, climate solutions, accessibility, healthcare, and more. There's so much meaningful work to be done. In closing, my best advice: build cool stuff with cool people while solving problems you're excited about solving. It's just that simple. Everything else will take care of itself. So what's next? You'll find out very soon but I could not be more excited. As always, Cheers :)
-
Reza Zadeh reacted on thisReza Zadeh reacted on thisthis is why I have trust issues... . I would have settled for 1 out of 4...
-
Reza Zadeh reacted on thisI’m excited to share that Gimlet Labs, Inc. has raised an $80M Series A, led by Menlo Ventures and joined by Eclipse, Factory, Prosperity7 Ventures, and Triatomic Capital. 5 months ago we launched with a thesis: AI inference is a fundamentally heterogeneous workload running today on homogeneous infrastructure, which leaves significant performance on the table. Agentic workloads are even more varied: chaining models, tool calls, code execution, and data retrieval. Gimlet orchestrates agentic inference across different silicon, slicing the workload and running each slice on the best hardware for the job. We’re seeing 3-10X speedups on large frontier models for the same power envelope. We’re also seeing exciting customer traction, and our customer base has tripled since launch and now includes a major model provider and hyperscaler. None of this would be possible without our team, which is the best I’ve ever worked with. Every breakthrough - from our compiler stack to the literal plumbing challenges in data centers - comes from this group. In addition my co-founders have been instrumental in getting us here, thanks Natalie Serrino, James Bartlett, Omid Azizi, and Michelle Nguyen! We’re also incredibly fortunate in the people backing us. Reimagining the entire inference stack requires high conviction, and we’re grateful to partner with an incredible set of people who share that conviction. Thank you to Tim Tully, Lior Susan, Chris Re, Abhishek Shukla, Jeff Huber, Willem van den Bosch, Derek Xiao, Sam Jackson, and our newest angels Lukas Biewald, David Kanter, Andrew Tan, Sharon Zhou, PhD, Bill Coughran and to the angels and advisors who have backed us from the beginning: Pete Warden, Dylan Field, Akshay Kothari, Amarjit Gill, Lip-Bu Tan, Andy Jacques, Raghu Raghuram, Mark Horowitz, Nick McKeown, Achin Bhowmik, Jonathan F., Roopak Venkatakrishnan, Karthik Narain, Kamal Shah, Shailendra Desai, Frank Serrino, Paul Sciarra, Dave Fowler, Vikram Chatterji. We're hiring across the stack: https://lnkd.in/gx8bgFHX Blog: https://lnkd.in/g9QCk22f TechCrunch by Julie Bort: https://lnkd.in/gS4j7P6eReza Zadeh reacted on thisWe're thrilled to announce Gimlet Labs' Series A funding, led by Menlo Ventures, and joined by Eclipse, Factory, Prosperity7 Ventures, and Triatomic Capital. Gimlet was founded out of the belief that AI inference would become the decade's defining infrastructure challenge. That challenge has become even more urgent with the rise of agentic workloads. Today, the top model providers are investing significantly in reducing agent latency. To deliver the next step of agent performance that will enable the next generation of software, we need to re-imagine the current infrastructure these agents run on. We designed Gimlet with heterogeneous hardware at its core, leveraging different types of hardware for different tasks, with a software orchestration layer to handle the complexity. So far, the result is 3-10X speedups on frontier models with large context windows, within the same power envelope. Read more here: https://lnkd.in/gxe9j6Vj
-
Reza Zadeh liked thisReza Zadeh liked thisRecently, I completed my internship position as a Supplier Quality Engineer at Mercedes-Benz U.S. International, Inc. I had the opportunity to work abroad, step out of my comfort zone and gain clarity on the type of engineer I aspire to be. During my internship, I utilized quality methodologies to diagnose root causes and implement counter measure for series production issues. I also supported Mercedes-Benz repair operations focused on cost saving initiatives and trained computer vision software Matroid to validate defects for inbound inspection. I am grateful for the opportunity to have worked at such an amazing organization, while growing both professionally and personally. A special thank you to my team in MP/NI3, my mentor Ingo Kutzer, and my fellow Interns Danielle Toland and Keren Judith Scherr for their guidance, support, and encouragement. One last thank you to my intern counterparts from all over the world, this chapter is not a goodbye, but a see you later. I look forward to reconnecting again in Germany :)
Experience & Education
-
Matroid
******* ***
-
******** **********
******* *********
-
**********
******** ****
-
******** **********
***** ************* *********** undefined
-
-
********** ** ********
********* ************ ******** ******** ******** ***********
-
View Reza’s full experience
See their title, tenure and more.
Welcome back
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
New to LinkedIn? Join now
or
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Licenses & Certifications
-
-
Commercial Aircraft Pilot. Instrument, Traildragger, Aerobatics, High Performance aircraft.
Federal Aviation Administration
Issued
Publications
-
MLlib: Machine Learning in Apache Spark
JMLR
Full list of authors:
Xiangrui Meng, Joseph Bradley, Burak Yavuz, Evan Sparks, Shivaram Venkataraman, Davies Liu, Jeremy Freeman, DB Tsai, Manish Amde, Sean Owen, Doris Xin, Reynold Xin, Michael J. Franklin, Reza Zadeh, Matei Zaharia, Ameet TalwalkarOther authorsSee publication -
Rapid estimate of ground shaking intensity by combining simple earthquake characteristics with Tweets
In proceedings: 10th National Conference on Earthquake Engineering
View Reza’s full profile
-
See who you know in common
-
Get introduced
-
Contact Reza directly
Other similar profiles
Explore more posts
-
NVIDIA AI
2M followers
Get faster and smarter MOE inference straight out of the box. 👇Deep dive on scaling expert parallelism with TensorRT-LLM. LLMs with MOE promise higher model capacity without linearly increasing compute costs - but they introduce new challenges -- more conditional computation, dynamic routing, and non-uniform GPU utilization -- solved by TensorRT-LLM. ✨ New in TensorRT-LLM has native support for expert parallelism—designed for fast, efficient inference with MoE models like Mixtral (Mistral AI) and DeepSeek (DeepSeek AI). This gives you: ✅ Dynamic expert routing: Automatically route tokens to the top-k experts with minimal overhead. ✅ Efficient expert scheduling: Balance expert loads across GPUs using smart sharding and token bucketization. ✅ Memory-aware execution: Maximize hardware utilization while respecting memory budgets. ✅ Drop-in support: Use @HuggingFace models with minimal code changes via TensorRT-LLM's #Python API. 🧠 How it works: MoE models activate only a subset of "experts" for each token. This dynamic nature is powerful—but hard to optimize. It’s all done under the hood using custom #CUDA kernels and NCCL-based communication primitives—giving you low latency, high throughput, and better GPU scaling. ✨ TensorRT-LLM handles: ✅Token-expert mapping using the gating network. ✅Token sorting to batch same-expert tokens together. ✅Expert parallel execution across GPUs. ✅Merging outputs for final predictions. 🛠️ Developer workflow - here is the code to get started. # Clone the repo git clone https://lnkd.in/g-GiDX23 # Use included examples to load and run a Mixtral model cd TensorRT-LLM/examples/mixtral From there, the Python API lets you load the model, convert it with TensorRT, and run expert parallel inference—all with a few lines of code. Results? 📈 Performance at scale. Tests show up to 2.3x faster inference throughput compared to standard tensor parallelism when using 8 GPUs and top-2 experts per token. Even better—TensorRT-LLM keeps efficiency high across increasing batch sizes. Want to see it in action or contribute? 👉 Read the full tech blog: https://lnkd.in/g_7YV3vV 👉 Explore the code on GitHub: https://lnkd.in/gNjQ5W2U 👉 Follow updates in the TensorRT-LLM repo: https://lnkd.in/gqSHYQ4u Share your experiences with us.
78
3 Comments -
Kameshwara Pavan Kumar Mantha
Equal AI | Your AI Call… • 7K followers
🚨 Two AI Legends Just Said the Same Thing: LLMs Aren't the Path to True Intelligence Richard Sutton (Father of Reinforcement Learning, 2024 Turing Award Winner) and Yann LeCun (Meta's Chief AI Scientist) are both moving beyond LLMs. Here's why this matters for everyone: 🧵 The Problem They Both See 🤔 Current AI chatbots (like ChatGPT) are basically sophisticated copy-cats. They: Mimic what humans would say ✍️ Don't actually understand consequences Can't truly learn from mistakes Have no real goals beyond predicting text Think of it like this: A parrot 🦜 can repeat "the sky is blue" perfectly, but it doesn't understand what sky, blue, or "is" means. What Real Intelligence Needs 🧠 Both experts say we need AI that: 1️⃣ Learns like a child does 👶 Not by copying, but by trying things and seeing what happens 2️⃣ Understands the physical world 🌍 Can predict: "If I push this cup, it will fall" 3️⃣ Has actual goals 🎯 Not just "predict the next word" but "achieve this specific outcome" 4️⃣ Plans ahead ♟️ Like chess players thinking multiple moves ahead, not just reacting The Future They Envision 🔮 Instead of bigger chatbots, imagine AI that: Learns continuously from experience 📈 Updates its understanding in real-time Can actually reason about cause and effect Builds mental models of how things work Why This Matters to You 💡 We're potentially at a turning point. The current AI boom might hit a wall 🧱, and the next breakthrough will come from completely different approaches. For businesses: Don't just bet on scaling current AI - watch for fundamentally new architectures For careers: Understanding real-world interaction and planning might become more valuable than prompt engineering The Bottom Line 📝 Two of AI's most influential thinkers are saying the same thing: We've been building really impressive calculators 🖩 when we should be building systems that actually understand the world. The next phase of AI won't just be ChatGPT-5 or GPT-6. It might be something entirely different. 🚀 What do you think - are we heading toward a major shift in how we approach AI? source: https://lnkd.in/gsHZvNbE https://lnkd.in/gchsKwsE #AI #GenAI #AGI #LLM #
41
3 Comments -
LangChain
502K followers
🧠💬 Memory in LLMs A practical guide showing how to implement conversational memory in LLMs using LangGraph, demonstrated through a therapy chatbot. Features code examples for basic retention, trimming, and summarization approaches. Learn to build memory-aware apps 👉 https://lnkd.in/gybcrV5v
968
21 Comments -
Ayush Gupta
Genloop • 6K followers
#TuesdayPaperThoughts Edition 60: The Art of Scaling RL Compute for LLMs This week's #TuesdayPaperThoughts explores "The Art of Scaling Reinforcement Learning Compute for LLMs" from researchers at Meta, The University of Texas at Austin, University of California, Berkeley, and Harvard University. While pre-training has had its moments with predictable scaling laws, RL training has remained more art than science until now. Key Takeaways: 1️⃣ Predictive Scaling Framework: The work introduces sigmoidal compute-performance curves for RL training that separate asymptotic performance (A) from compute efficiency (B). This framework enables extrapolation from smaller-scale runs to predict performance at larger compute budgets. The team validated this with a massive 100,000 GPU-hour run where predictions from just the first 50k hours closely matched final performance. 2️⃣ Not All Recipes Scale Equally: Methods that look promising at small compute budgets can hit lower performance ceilings at scale. The study reveals that design choices like loss type and FP32 precision shift asymptotic performance, while factors such as loss aggregation and normalization primarily modulate compute efficiency. This explains why some widely-adopted methods plateau unexpectedly. 3️⃣ ScaleRL Recipe: Through systematic ablations consuming 400,000+ GPU-hours, the team developed ScaleRL—combining PipelineRL with 8-step off-policyness, truncated importance sampling (CISPO), FP32 logits computation, and adaptive prompt filtering. ScaleRL achieves A=0.61 asymptotic performance, outperforming DeepSeek's GRPO, Qwen's DAPO, and other prevalent methods on both ceiling and efficiency. The timing couldn't be better. With RL compute budgets exploding (10× increase between model generations for o1→o3 and Grok-3→Grok-4), the field desperately needed this systematic approach. Research Credits: Devvrit K., Lovish Madaan, Rishabh Tiwari, Rachit Bansal, Sai Surya Duvvuri, Manzil Zaheer, Inderjit Dhillon, David Brandfonbrener, Rishabh Agrawal Paper Link: In comments
21
1 Comment -
Ayan Sinha
Upwork • 2K followers
I’m excited to attend NeurIPS workshop sessions at San Diego and present two papers authored by the applied research team: 1. GraphMatch: Fusing Language and Graph Representations in a Dynamic Two-Sided Work Marketplace. Mikołaj Sacha et al. - NeurIPS Workshop on Unifying Representations in Neural Models (UniReps) 2. Towards Real-World Evaluation of Agentic Work in Freelance Marketplaces. Mattie Terzolo et al. - NeurIPS Workshop on LLM Evaluation Over the next couple of days, I’ll be sharing insights and some behind-the-scenes perspectives on how we think about these models, how they’re shaping Upwork today, and the opportunities we see ahead. For attendees already onsite at the main conference, please stop by the Upwork booth #722 to learn more about the efforts the ML & AI team is driving - from research to real product impact. Feels great to be again publishing at NeurIPS after a hiatus and looking forward to great conversations and inspiration ahead!
50
2 Comments -
CoreWeave
118K followers
Today marks another step toward making Reinforcement Learning (RL) faster, simpler, and scalable for builders harnessing the Essential Cloud for AI. Our support on the development of torchforge - a new PyTorch-native framework - is helping researchers take large-scale RL workloads from experimentation to production. In partnership with Meta and Stanford University, torchforge was validated on CoreWeave’s Cloud Platform proving stable, efficient, and highly performant – ready to power the next generation of RL workloads. torchforge is now supported on SUNK. More in our recent blog: https://hubs.la/Q03PLJC10
136
-
Sharon Zhou, PhD
AMD • 43K followers
Excited to share our NeurIPS 2025 Tutorial on How to Build Agents to Generate Kernels for Faster LLMs (and Other Models!) A collaboration across institutions: AMD, Stanford, Google DeepMind, Arm, Nvidia, Meta, Modular AI, UC Irvine, and ML Commons. - If you're an AI researcher, check out the parts of AI compute (e.g. GPU) that affect your inference & training jobs -- from why you should do batch sizes in certain multiples of 2 to how your trillions of matrix multiplications make their way through different types of memory on a chip - Also chat with us about the trend of self-improving LLMs that can write code to make them even faster. Agents -> SFT data -> RL with profiling tools for strong reward signals - Come build with us! It's an exciting future where LLMs can make the next generation of AI research (RL, even more sparsity, hybrid models with SSMs) super fast on any new hardware generation -- on-the-fly in your IDE Sina Rafati, Hao Li, Azalia Mirhoseini, Anna Goldie, Laurence Moroney, Vartika Singh, Mark Saroufim, Chris Lattner, Sitao Huang, David Kanter, Simon Guo, Kesavan Ramakrishnan, Vincent Ouyang, Tim Gianitsos, Nithyashree Manohar, Sharon Zhou, with acknowledgements to the AMD GEAK team Zicheng Liu, Dong Li, Ziqiong Liu, Pratik Prabhanjan Brahma and the AMD ROCm Kernel team ZHAOYI LI and the AMD Omni team Muhammad Awad, Keith Lowery, Cole Ramos and kernel engineers John Tyler, Muhammad Osama and of course the one and only Emad Barsoum. When they're out, will drop... * arxiv paper link * Github repo link
1,298
39 Comments -
Matthew Collier
TOBY • 5K followers
Yann LeCun was recently covered on X saying... "I'm not interested in LLMs anymore - they're the past. The future is in four more interesting areas: machines that understand the physical world, persistent memory, reasoning, and planning." (https://lnkd.in/gX2JdhCY). Why would he make such a statement? The reason lies in the fact that while massive investment in dollars, compute, energy and innovation continues to grow unabated in AI, there remain a nagging set of persistent questions. These questions pertain to the intersection of the newly evolving world of intelligent machines and the practical, deterministic physical reality in which we all live. To reach AI-full potential we will need to create an orchestrating interface between these intelligent machines and the real-world and a collaborating fabric between the multitude of intelligent machines themselves. The attributes that will define such a reliable and productive network are as follows: 1.) The network itself will need to be intelligent and adaptive; able to handle high levels of complexity at global scale and at the edge with reasoning that allows for real-world disruptions that are endemic to everyday life. 2.) The network will also need to be informed by and anchored to the intent of the people guiding it with the planning capability to extend those human derived objectives into action and follow-through. 3.) The architecture will need to be persistent as the machines will need to be capable of autonomy; knowing what they were doing, what they are doing and what they intend to do. 4.) The machines need to be capable of coordinating amongst themselves. With 5.3 Billion people now on the internet and more than 400 Billion machines the future of the internet will be not just about human-to-machine communication but primarily about the machine-to-machine collaboration. 5.) The intelligence will need to be deterministic. Gen-AI with it's probabilistic logic will only get you so far in orchestrating real-time, real world actions that, by definition, are deterministic. 6.) Finally, an aspect not mentioned by Yann LeCun is the increasing importance of cyber-security in a world in which the Intelligence is on the move and actively engaging with people "in the wild". Welcome to the Paranet; a secure, distributed intelligence, asynchronous compute that leverages non-stochastic AI and persistence to create "the Internet of work" where humans and machines but primarily machines to machines can collaborate securely and reliably at the edge.
15
2 Comments -
Jessica Hawk
25K followers
Customers are asking for AI that goes beyond assistance to true collaboration. Claude Opus 4.5 from Anthropic delivers on that vision: understanding objectives, factoring in constraints, and ready to power complex workflows. Opus 4.5 joins the full set of Claude models now in Microsoft Foundry, alongside models from OpenAI, Mistral AI, Cohere, Meta, and others, so you can easily compare and find the best-fit models for your needs. With Microsoft Foundry, Azure customers get the broadest selection of model providers on any cloud, plus the governance, security, and interoperability to accelerate innovation with confidence. 🔗 https://lnkd.in/gdRxNB_q
107
3 Comments -
Furu Wei
Microsoft Research Asia • 12K followers
Introducing Generative Adversarial Distillation (GAD): a novel GAN-style formulation and framework that facilitates both on-policy and black-box distillation of large language models (LLMs). GAD is the first technique to enable block-box on-policy distillation from proprietary teachers where internal logits or parameters are inaccessible, or distillation between teacher and student LLMs with incompatible vocabularies. GAD expands our prior work on white-box on-policy distillation (i.e., MiniLLM), pioneering block-box on-policy distillation for LLM training. Specifically, GAD frames the student LLM as a generator and trains a discriminator to distinguish its responses from the teacher LLM’s, creating a minimax game. The discriminator acts as an on-policy reward model that co-evolves with the student, providing stable, adaptive feedback. Experimental results show that GAD consistently surpasses the commonly used sequence-level knowledge distillation. In particular, Qwen2.5-14B-Instruct (student) trained with GAD becomes comparable to its teacher, GPT-5-Chat, on the LMSYS-Chat automatic evaluation. The results establish GAD as a promising and effective paradigm for black-box LLM distillation. Our team has been conducting fundamental research in knowledge distillation with wide adoptions across the industry. - MiniLM: We introduced multi-head attention distillation, establishing the most effective distillation method for BERT-style models. The open-source MiniLM models (e.g., 6x384) have become the most widely utilized small encoder models on the Hugging Face. - MiniLLM: Our proposed Reverse KLD is recognized as one of the most effective, de facto on-policy distillation approaches for modern LLM training, which has been widely used by Thinking Machines, Gemma, and many other teams and models. - BitDistill: We proposed BitNet Distillation to finetune off-the-shelf full-precision LLMs (e.g., Qwen) into 1.58-bit precision (ternary weights {-1, 0, 1}), achieving performance parity with the full-precision counterparts on specific downstream tasks. - GAD: The development of Generative Adversarial Distillation (GAD) now allows for black-box on-policy distillation, overcoming two major prior limitations: (1) Distillation from proprietary teachers where internal logits or parameters are inaccessible; (2) Distillation between teacher and student LLMs with incompatible vocabularies. https://lnkd.in/gMaP2c7w
125
2 Comments -
Eitan Anzenberg, PhD
Eightfold • 3K followers
Out team just posted our latest paper “Evaluating the Promise & Pitfalls of LLMs in Hiring Decisions” on arXiv! We found some exciting results: • Benchmarked leading LLMs (GPT-4o, o3, Claude, Gemini, Llama, DeepSeek) against Eightfold’s “Match Score” model on real-world data. • Evaluated both performance (ROC AUC, PR AUC, F1) and fairness (impact-ratio across gender, race, intersectional groups). • Eightfold’s Match Score beat the best LLM on accuracy (ROC AUC 0.85 vs 0.77) and fairness (min race Impact Ratio 0.957 vs 0.809). • Off-the-shelf LLMs still propagate measurable demographic bias without safeguards. • The trade-off between accuracy and fairness is a false dichotomy: carefully engineered, domain-tuned models like Eightfold’s can achieve both accuracy of hiring and fairness of outcomes. https://lnkd.in/guQ2TAYp #machinelearning #ai #eightfold #arxiv #datascience #bias #fairness #ml #data #genai #llms
37
2 Comments -
John Gibbon
Further Advisory - a Tredence… • 4K followers
GenAI is Changing the Game – Devs, You’re Evolving. PMs, You’re Lagging I recently attended SF Awesome AI Dev Tools hosted by Yujian Tang https://lu.ma/x5v9zmo6 at GitHub - my favorite venue for AI meetups. My favorite presentation of the evening was from Jules Damji from Databricks; huge thanks for letting me share his deck on how software engineering is adapting to the #GenAI revolution: https://lnkd.in/gKUmrFxp ⚡️AI is Changing Daily Work We’ve all seen how GenAI is changing the way we work — from tools like ChatGPT to copilots built into productivity apps. But the impact goes far beyond that: #GenAI is reshaping customer and employee support, sales and marketing content creation, data analytics, and of course, software development. 🤖 AI is Changing Software Development We’ve all heard the hype; but Jules is on the front lines and his insights really stood out. TL;DR · AI is changing the way we will code · Collaboration with AI is the new way forward · We are in this early .. things are changing rapidly · Adapt and thrive in the new age 📈 AI Code Gen is Taking Off - Fast · Satya Nadella recently noted, “AI writes up to 30% of Microsoft’s code” · An Anthropic engineer claimed "90% of the Claude Code is written with Claude Code” · The creator of Claude Code added: "We are only scratching the surface of what's possible. The model is moving really fast -- it's exponential and it's getting better at coding very, very, quickly” ⚠️ Product Managers Beware! At Y Combinator’s AI Start-up School, Andrew Ng recently said that in AI start-ups, product management is now the bottleneck – not engineering. Watch his "Building Faster with AI" talk: https://lnkd.in/gya6uDd8 Product managers, we need to step up our #GenAI game! (Blog post to follow.) (Views are my own and do not represent those of any current or former employer.) #GenerativeAI #GenAI #SoftwareDevelopment #ProductManagement #ArtificialIntelligence #FutureOfWork #AIProductivity #AIProductManagement
35
3 Comments
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top contentOthers named Reza Zadeh
305 others named Reza Zadeh are on LinkedIn
See others named Reza Zadeh