Lauren Tan, known online as poteto, builds developer tools and works on the React Compiler. The React team lists her as an engineer at Cursor. Her work spans compiler optimization, engineering leadership and practical workflows for coding agents.[1][2][3]
Start with pstack →
X Background
Tan began her technology career as a user-interface designer and learned programming to bring those designs to life. Her official React biography describes studying finance and learning to code through Excel. That path informs her writing for engineers who enter the profession without a computer-science background.[1][5]
Career
Tan previously managed engineers at Netflix and worked at Facebook, later Meta. In February 2020 she announced that she was leaving Netflix management to return to an individual engineering role. Her current React team biography lists Cursor, while her GitHub describes work at xAI and membership of the React Compiler core team. The verified continuity is her work on developer tools; the public pages use different employer labels.[1][2][5][6]
Contributions
Making React optimization automatic
Tan co-authored the React Compiler 1.0 announcement with Joe Savona and Mofei Zhang. The compiler analyzes how values flow through React components and hooks, then adds memoization so an application can reuse calculations instead of repeating them unnecessarily. The release account credits her with work on a major compiler rewrite. This reduces the amount of manual optimization an application developer has to maintain, while still requiring code to follow React’s rules.[3]
Hiring and learning in public
Tan’s hiring-without-whiteboards repository collects companies whose interview processes avoid conventional whiteboard exercises. It is a community-maintained resource for finding alternatives, rather than a guarantee about any employer’s current interview. Her blog complements that project with writing on leadership, communication and becoming more capable through unfamiliar work.[7][5]
Ideas
Tan wrote pstack to turn her engineering practices into instructions that coding agents can follow. Its skills divide work into planning, implementation, verification and review, with explicit principles for deciding what to build. In the README she argues that higher code output is not the objective: confidence in parallel agents depends on verifiable results. Cursor’s Compile events feature her presenting this style of work; the activity counts in a talk title are not a general productivity benchmark.[8][11][10]
Tools, writing & talks
Tools
Start here
pstackcursor.comA Cursor plugin with structured agent skills for planning, implementation and review. Read its workflow before enabling automation in a repository.
Start with pstack’s linked README and guide to understand how Tan turns engineering principles into agent workflows. Read one verification skill before applying it to a small repository; familiarity with a coding agent helps.
- hiring-without-whiteboards
github.comA community-maintained directory of employers with alternatives to traditional whiteboard interviews.
Repository owner[7]
Writing, research & talks
Part 2 of a pstack guide on agent-driven research, planning and architecture. Covers indirect prompts, history-aware context recall, README-driven design, parallel prototyping, and a cross-model architecture arena, all anchored on empirical verification rather than abstract plans.
Catalog workOriginal sourcepstack v0.15.27 adds a beta meta-skill that builds formal verification skills for agents, covering TLA+ modeling and Kani checks for Rust. The author explains the approach and reports bugs found in their own code. Thread collection was partial.
Catalog workOriginal sourceThe pstack author explains how 24 concise principle skills work: a mode skill loads an index, agents must name each principle applied and the decision it changed, and users steer with principle names. He also argues much of this will compress as models improve.
Catalog workOriginal sourceArgues that self-verification is the foundation of an agentic loop: agents exercise the real interface and produce videos, screenshots or formal models as reviewable proof, with honest caveats about approximation.
Catalog workOriginal sourceA walkthrough of automating release and QA with two team bots: contributor notifications, a fuzz swarm of 10+ agents testing the build, and an engineer bot that triages, fixes and cherry-picks patches.
Catalog workOriginal source- React Compiler v1.0react.dev
The release announcement explains automatic memoization, the compiler architecture and how teams can adopt it incrementally.
Co-author and compiler contributor[3] - Does It Spark Joy?no.lol
Tan explains why she left engineering management at Netflix to return to individual engineering work.
Author[6]