Lauren Tan

Lauren Tan, known online as poteto, is a software engineer at xAI and a member of the React Compiler core team. Her work spans compiler optimization, interface development, and workflows for coding agents. She helped develop React Compiler, which reduces the performance bookkeeping React developers do by hand, and created pstack, which packages her approach to investigating, implementing, and checking agent-assisted changes.

Moving between engineering and management

Lauren grew up in Singapore and began exploring web design as a teenager through tutorials, viewing page source, and online design communities. She later designed websites for schools and small businesses. She studied finance, learned programming with Excel, and entered technology through UI design before becoming a software engineer.

At Netflix, Lauren led engineering for the Studio Programming team, building software for scheduling and programming original content. She subsequently managed engineers in the UI Engineering organization working on the company’s content-creation studio. These tools served the people making Netflix productions.

In Does It Spark Joy?, Lauren describes entering management after the departure of a manager who had helped her develop as a technical leader. She accepted the responsibility as a choice she could revisit. After two years, she found that the rewards of coaching people and building a team did not replace the pleasure of making things herself. She announced that she would leave Netflix in February 2020, after four years, to return to hands-on engineering.

At Meta, Lauren led internal prototyping of React Server Components and added support for them to Relay’s compiler and runtime. She later became an engineering manager supporting React. Her career has included both technical implementation and responsibility for the people developing the framework.

Making React performance less manual

React Compiler analyzes JavaScript during the build and automatically reuses computed values when appropriate, a technique called memoization. This reduces the need for developers to maintain performance annotations themselves.

Lauren worked with Joe Savona, Sathya Gunasekaran, and Mofei Zhang on an architectural rewrite that represented programs using a control-flow graph—an explicit account of the paths execution can take. That representation enabled more precise analysis and type inference. The rewrite was part of a sustained team effort, informed by earlier prototypes and subsequent revisions.

The compiler can, for example, reason about values after a conditional early return, where manually placing a memoization hook would violate React’s rules. Its analysis also powers lint diagnostics that can expose bugs. These capabilities explain why the project goes beyond inserting familiar optimization calls: it can analyze program structure that developers cannot express through those hooks alone.

Lauren authored the October 2024 beta announcement, which introduced support for applications using React 17 and later and opened the compiler’s working group to broader participation. She coauthored the October 2025 stable-release announcement with Joe and Mofei. Version 1.0 brought automatic optimization to both React and React Native.

Giving coding agents a working method

Lauren’s pstack packages her engineering workflows for coding agents, with an emphasis on less code and stronger verification rather than maximizing output. She introduced and merged the plugin into Cursor’s public plugin repository in May 2026.

Its routing skill, poteto-mode, selects a playbook suited to the task—such as a bug fix, feature, performance investigation, or plan—and calls supporting skills as needed. This makes the workflow responsive to the kind of engineering work underway rather than prescribing the same sequence for every request.

The verification workflow makes the checking process concrete. It generates project-specific instructions for launching an application, exercising it as a user would, and collecting observable results. A check should cover relevant side effects, such as whether data persisted, as well as what appeared on screen. The instructions must be usable by a fresh agent unfamiliar with the application.

Practices shaped by experience

Lauren’s writing and community projects explain several positions behind her engineering work:

• Choose techniques for the problem. In This Time Is Different, Lauren revisits her earlier confidence that adopting the right patterns could future-proof a codebase. Learning Elixir challenged habits formed in Ruby and JavaScript. Functional techniques expanded her options, but did not make a straightforward loop obsolete. She asks engineers to weigh the problem, opportunity cost, and tradeoffs their team can accept.
• Make the team more effective. Her advice on becoming a tech lead emphasizes organization, communication, and helping others deliver. She developed that advice through one-on-one conversations offered to people from underrepresented groups. Technical leadership does not require being the strongest programmer in every room.
• Interview for the work people will do. Lauren maintains Hiring Without Whiteboards, a community-contributed directory of employers using practical engineering interviews. It distinguishes discussion of real problems from trivia and puzzles, and includes approaches such as pairing sessions and work samples that let candidates demonstrate abilities relevant to everyday engineering.