Simon is the Co-founder of Inferact, a startup founded by creators and core maintainers of vLLM, the most popular open-source LLM inference engine. Inferact's mission is to grow vLLM as the world's AI inference engine and accelerate AI progress by making inference cheaper and faster. Simon has co-led the vLLM community effort from UC Berkeley since 2023. Simon previously worked at Anyscale and Character AI on serving systems.

Katelyn Lesse is the Head of Platform Engineering at Anthropic. She leads engineering for the Claude Platform, including APIs and developer tooling, as well as Anthropic’s product infrastructure. She has experience building and scaling engineering organizations across fintech and developer platforms, with previous leadership roles at Stripe and Betterment.

Dmytro "Dima" Dzhulgakov is the CTO and Co-founder of Fireworks, the platform for specialized intelligence enabling companies like Uber and Shopify to train and serve custom models. Before Fireworks, Dima served on the leadership team at Meta's PyTorch for over a decade.

Nikola Borisov is CEO and Co-founder of DeepInfra, an AI inference cloud platform democratizing access to AI by bringing the latest open-source models into production. DeepInfra's mission is to give businesses control over their AI stack through scalable, cost-efficient inference infrastructure

Jun Qian is Chief AI Officer at GMI Cloud, leading inference and agent engineering. He has 25+ years in AI and machine learning, with engineering leadership roles at Oracle, Microsoft and Amazon, most recently as VP of Generative AI Services at Oracle.

Natalie Serrino is a co-founder of Gimlet Labs and leads kforge, Gimlet’s kernel generation system for diverse accelerators. She applies her operational background to heterogeneous inference orchestration: assigning individual stages of agentic applications to processors suited to their distinct computational and memory requirements. Prior to Gimlet Labs, she was a founding engineer at Pixie Labs. Following Pixie’s acquisition by New Relic in 2020, she worked there as a principal engineer, developing infrastructure observability techniques including Kubernetes autoscaling based on application-specific HTTP traffic. She studied computer engineering at Brown University, where she helped establish Spira, an engineering program for high-school girls supported by the National Science Foundation.

Banghua is co-founder and CTO of RadixArk, where he is building open infrastructure for frontier AI inference and post-training. The team at RadixArk works on SGLang, the high-performance serving engine for large language and multimodal models, and Miles, RL framework for post-training at frontier scale. Prior to RadixArk, he was an assistant professor at the University of Washington and a principal researcher at NVIDIA, and earned his Ph.D. from UC Berkeley after completing his undergraduate studies at Tsinghua University.

Ryan Hanrui Wang is SVP, Applied AI at Nebius, where he leads inference, post-training, and agentic systems. He previously co-founded and served as CEO of Eigen AI, acquired by Nebius in 2026. Under his leadership, Eigen AI built a full-stack inference platform that achieved leading performance across 25 major open-source models and was recognized as the #1 speed inference provider during NVIDIA CEO Jensen Huang’s GTC 2026 keynote. Ryan earned his Ph.D. in Computer Science from MIT as Professor Song Han’s first doctoral student. His research spans efficient AI computing, generative AI inference, and computer architecture. He developed SpAtten, a pioneering sparse-attention framework introducing cascaded KV cache pruning and quantization, and is the most-cited HPCA paper published since 2020. He also introduced Hardware-Aware Transformers. His open-source models and repositories have surpassed one million downloads and been incorporated into IBM products and the PyTorch Ecosystem. Ryan received the MIT Technology Review Innovators Under 35 award, Best Paper Awards at QCE and ICML RL4RL, first place in the ACM Student Research Competition, Qualcomm and Unitary Fund fellowships, and Rising Star recognition in ML & Systems and ISSCC. He co-founded the QuCS Forum to promote AI education and earned his B.Eng. with highest honors from Fudan University.

Nancy Zheng is the VP of Engineering of AI Inference at Cerebras. She's building the inference serving architecture for the fastest AI in the world. She worked on enabling OpenAI's GPT 5.6 Sol Ultrafast on Cerebras (15x faster!), and now she's building one of the world's first disaggregated inference clouds combining GPUs and Cerebras to get 5x throughput. Previously, Nancy was a founder and led teams at Google and Stripe.

Omid Azizi is a co-founder of Gimlet Labs, and head of hardware platforms. His work focuses on enabling disaggregated inference across current and emerging hardware, pushing the limits of inference performance. Prior to Gimlet Labs, Omid was a founding engineer at Pixie Labs, where he developed novel approaches to software observability. Following Pixie’s acquisition by New Relic, he led the company’s Kubernetes observability efforts. Earlier, he was a founding engineer at Hicamp, later acquired by Intel, where he worked on server CPUs and AI training chips. Omid received his PhD in Electrical Engineering from Stanford.

Harsh Banwait is Senior Director of Product at CoreWeave, where he leads product strategy for the CPU and GPU platforms powering the next generation of AI. He partners with leading AI labs and technology companies to bring cutting-edge compute platforms to market, shaping the cloud infrastructure that turns ambitious AI research into reality at scale.

Sami Jaghouar is Head of Research at Prime Intellect, where he leads work on large-scale reinforcement learning and post-training infrastructure for frontier AI models. He created and leads PRIME-RL, an open-source asynchronous reinforcement learning framework designed to scale reasoning and agentic training across hundreds to thousands of GPUs, and was a core researcher behind INTELLECT-3, a 106B-parameter mixture-of-experts model that achieved state-of-the-art performance for its size. His earlier work includes leading the INTELLECT model series and creating OpenDiLoCo, an open-source framework for communication-efficient distributed training.

Charles builds and teaches people to build inference-intensive applications. He got his PhD from the University of California, Berkeley, in 2020 for work on the geometry of neural network optimization. He has since worked as an engineer of and educator about inference applications at Weights & Biases, Full Stack Deep Learning, and now Modal

Nick Khami leads inference at Morph, which serves open models tuned for coding agents. He runs Morph's serving stack for models like Kimi K3 and GLM on GB300 clusters, using NVIDIA Dynamo with disaggregated prefill and decode, and works on KV cache routing and reuse across the fleet. Before Morph, Nick founded and led Trieve (YC W24), a search and RAG infrastructure company that a16z-backed series B startup Mintlify acquired, and later led engineering at Mintlify. He has spoken at Code with Claude and Vercel AI, and writes about inference engineering at skeptrune.com (opens in a new tab). Nick has a B.S. in Computer Science from UT Austin.