People

Roboflow’s cofounder and CEO builds tools that connect computer vision models to useful applications, from correcting training labels to deploying systems that act on camera feeds.

Singapore’s digital development minister, whose AI policy work spans workforce training, practical assurance and public-sector adoption.

Imbue co-founder and CTO whose work spans cost-aware model training, infrastructure reliability, and public views on coding agents and uncertainty.

AI researcher whose documented work spans post-training at OpenAI and data-centric error analysis.

Creator of Transformers.js, bringing pretrained machine-learning models into browsers and JavaScript applications.

Sales and portfolio operator who cofounded Roadrunner to rebuild enterprise quoting around modern pricing and AI-assisted deal workflows.

Elicit cofounder and COO whose work centers on structured AI research workflows, human evaluation, and rapid product iteration.

AI researcher whose work spans Qwen, multimodal pretraining, and the training systems behind tool-using agents.

Computer-vision researcher and World Labs cofounder whose collaborative work includes dense image captioning and perceptual losses for fast image transformation.

AI investor at Andreessen Horowitz whose public work explores creative workflows, personal AI companions, and software creation for consumers.

Product engineering leader at Cartesia whose work spans real-time AI audio delivery, coding-agent infrastructure, and interfaces for crafting text with LLMs.

Runway chief technology officer working on the infrastructure and evaluation of real-time generative world models.

Imbue’s CEO and co-founder connects dependable AI agents with a broader ambition: software that people can understand, modify and control.

Nous Research co-founder whose work includes the original Hermes collaboration and WorldSim’s experiments with language-model-generated environments.

Composio co-founder and CTO working on how AI agents discover, execute and improve tools, with an emphasis on reusable skills and human control.

AI researcher working across post-training, model evaluation and the design of human–AI collaboration, including ChatGPT Canvas and PostTrainBench.

Builder of Lightpage and Feedpaper, and an advocate for evaluating LLM applications through complete pipeline replays.

An engineer and product leader whose work on Windsurf and Google Antigravity centers on agents that can follow, continue, and verify software engineering work.

AI researcher at Letta and coauthor of sleep-time compute and MemGPT, two approaches to using persistent context in language-model systems.

First author of a NeurIPS 2025 Best Paper on scaling self-supervised reinforcement learning to thousand-layer networks; identified by Delta Podcast in May 2026 as an OpenAI researcher working on RL and reasoning.

Former OpenAI Chief Product Officer and science leader whose work included starting OpenAI for Science and leading the scientific workspace Prism.

Google DeepMind leader and Google Chief AI Architect whose collaborative research includes Atari deep reinforcement learning and WaveNet.

Daily cofounder and CEO working on Pipecat, voice-agent evaluation, and the infrastructure for real-time AI conversations.

OpenPipe co-founder whose work spans model customization, reinforcement learning for agents, and Y Combinator’s Startup School.

GitHub executive and former platform engineering leader whose work spans internal AI adoption, developer tools and the social trust surrounding agent-generated code.

NVIDIA engineer and Dynamo architect working on distributed inference, model performance, and the systems behind agent workflows.

Game developer and toolmaker whose AI work spans visual exploration interfaces, production coding experiments, and arguments about the future of games and software.

AI engineer and technical writer whose work examines agent context management and how application designs should change as models improve.

Developer, npm Inc. cofounder and writer on AI application design, evaluation and the changing work of software engineers.

Coauthor of OBELICS and the Idefics vision-language models, with research focused on open multimodal data and experimentally justified model design.