Connects Stripe's transaction data with payment foundation models, fraud prevention and authorization. Her technical accounts show how AI becomes part of the core product and the company's daily work.
Operators building and applying AI in products and company workflows. Selected for documented work, including past roles.
Cosign
An operator focused on the infrastructure around AI: evaluation, deployment and dependable search. His writing draws on systems he has built and teams he leads.
David is building exe.dev as infrastructure for developers and coding agents after co-founding Tailscale. His writing and interviews explain how prototypes, customer pain, and discarded experiments shape infrastructure choices for agent-driven software.
Built Replicate's AI model deployment platform
Makes AI evaluation concrete. Drawing on machine-learning work at GitHub and Airbnb, Hamel shares practical methods and open-source tools for understanding where AI systems fail and how to improve them.
As Stitch Fix’s founding Chief Algorithms Officer and a former Netflix data-science leader, Eric worked where machine learning met the operating model of the business. His writing makes a strong case for data scientists who own problems end to end, pairing technical platforms with the freedom to discover useful business ideas.
Romain leads developer experience at OpenAI, with practical accounts of Codex design, developer feedback, and agent delegation. His earlier Stripe platform work adds a useful perspective on turning powerful AI capabilities into tools developers can actually adopt.
Built OpenAI's Deep Research and agent systems
Brings engineering experience from Meta, Lyft and GoFundMe to his writing in Path to Staff Engineer. Now building around software and AI and exploring robotics—a useful perspective on growing your craft while continuing to build.
Doug brings search leadership from Reddit and Shopify to one of AI’s hardest practical problems: retrieving the right information. Tal Raviv recommends him for agentic search; his consulting and teaching connect relevance engineering with systems teams can actually build.
Paul explains how AI agents, human copilots, and a shared knowledge layer change the design of customer service. His Fin product leadership and detailed operating accounts make him useful for thinking through AI adoption as a product and workflow redesign.
An AI engineer who turned structured outputs into a widely used developer tool. His current Codex work builds on hands-on RAG, evaluations and consulting experience.
Mario created pi, the extensible coding agent he now develops at Earendil, after years building open-source tools including libGDX. Tal Raviv singles him out for exploring agents in public—a useful window into the design decisions behind tools other builders can extend.
Independent open source developer building tools for data journalism, SQLite, and large language models
Simon builds Datasette and LLM tooling and publishes detailed experiments with coding agents, model behavior, and AI security. His habit of shipping small tools and explaining their limitations makes his work useful to engineers deciding how to apply AI in practice.
A practitioner who pairs production ML experience with working prototypes. Useful for recommendations, evaluations and the engineering tradeoffs behind AI products.
Vicki built infrastructure for early OpenAI researchers and later co-founded Gantry, bringing a systems perspective to machine-learning teams. Her published work on research infrastructure and Kubernetes makes the operational choices behind AI development unusually concrete.
Chip turns production ML experience at NVIDIA, Snorkel AI, and Netflix into practical guidance on AI system design. Her books and open-source tools connect model capabilities to the evaluation, data, and engineering choices required to make them useful.
