Hello, Anthropic!

I'm Zeke. I am a designer and engineer and the first employee at Replicate, a platform for running AI models in the cloud. I've been at Replicate for four years now, and I'm starting to look for new opportunities. I'm interested in engineering roles at Anthropic, where I can help build the future of natural language programming.

I've worked on developer tools for more than 15 years at companies like Heroku, npm, and GitHub. I co-created Swagger (now OpenAPI), a specification that has become an industry standard and an integral part of Anthropic's Model Context Protocol. You can read my CV for a full summary of my work history.

My work is about making it easier and more fun for humans to use computers. I love building tools that help people achieve their creative goals. The recent rise of large language models with access to tools has been the most exciting development in my career, and is exactly the kind of leap forward that I've always wanted as a designer and developer.

I'm very excited about what you all are building, and I would love to be a part it.

~ Zeke

My recent work at Replicate and GitHub

I'm a generalist. I've worked on many things at Replicate, from product design to engineering to developer advocacy. I helped build our Django web frontend, our Golang HTTP API, and our usage-based billing system. I maintain our Python and TypeScript client libraries and our MCP server. I write documentation, blog posts, and changelog entries. I create video tutorials and open-source demo apps. I helped design and create Cog, an open-source tool for building and packaging AI models in Docker containers, with a standardized API interface based on OpenAPI, JSON Schema, Pydantic, and FastAPI.

Before joining Replicate, I was a staff engineer at GitHub for five years. I helped launch Electron, the open-source tool for building desktop apps, which has become the basis for apps like Slack, VS Code, Cursor, and Claude Desktop. On the Electron team, I internationalized electronjs.org, added Typescript support, and helped manage an active community of international open-source contributors, and created a governance model for the project. After Electron, I founded GitHub's first docs engineering team. We launched docs.github.com, internationalized docs.github.com, open-sourced docs.github.com, and co-created the OpenAPI description for GitHub's REST API.

I have extensive knowledge of the AI/ML landscape, having spent the last four years working full-time at an AI startup, and having had access to the earliest versions of Copilot while working at GitHub. Replicate runs thousands of AI models in production, from open-source models like Flux, Qwen, and Wan to proprietary models like GPT, Claude, Gemini, Veo, and others. I know the commonalities and differences between these models, and I've worked on tools, documentation, and demos to support users working with all AI modalities, from text to images to audio to video.

Claude

I use Claude Code every day, and it has become my favorite tool for writing code and automating tasks. I've been coaching my team at Replicate on best practices for using Claude, and we are starting to connect and adapt our many disparate codebases together in unforeseen ways. We experiment with lots of coding products like OpenAI Codex, Gemini CLI, Cursor, Warp, and others, but Claude Desktop, Claude Code CLI, Claude Code for GitHub, and Claude Code on mobile have taken the lead as our company's tools of choice. We're experimenting with all the various ways to bring Claude into all of our processes, not just for writing code, but for understanding our data in Metabase, our conversations in Slack, our internal documentation in Notion, and our project management in Linear.

Below you'll find some recent examples of things I've been working on with Claude Code, both in my professional and personal life.