Lauren Tan: Engineer Behind Grok Bot and React Compiler

Updated 2026-10-12Needs owner review

Needs owner review. Every statement below comes from a public source linked in the Sources section. No private details are included. Please re-check the sources before publishing, since job titles and projects change.

Who she is

Lauren Tan is a software engineer whose public profile says she works on Grok Bot and Cursor at SpaceXAI, with a current title listed as Principal Engineer. [1] Her LinkedIn profile describes the company as xAI, and her GitHub profile lists her as a software engineer at @xai-org and a member of the React Compiler core team. [1][2] She posts publicly as @poteto on GitHub and X. [2][6]

Background

According to her public profile, she previously worked at Meta and Netflix, and she is on the React Core team, working on React Compiler. [1] Her LinkedIn profile describes contributions to React Compiler and to the open-source release, and lists a prior role as a Member of Technical Staff at Cursor, where her profile says she “Shipped Cursor 3 and Grok Bot.” [1]

Work on agents and Grok Bot

Tan has spoken publicly about how she uses AI agents in software engineering:

• In an August 2026 LinkedIn post she wrote about “building pstack,” “shipping 2,000 PRs a month,” and “the client framework powering Grok Bot.” [3]
• A recorded talk and a written breakdown of it describe her approach: using architecture, CI, and verification steps so that agents can be trusted with more of the work, and coordinating several agents more like a manager. [4][5]
• The breakdown quotes her as saying, “I almost never look at code anymore,” which it attributes to architectural constraints, CI, and verification flows doing checks she previously did line by line. [5]

Open-source work

Her GitHub profile lists public repositories including hiring-without-whiteboards (a widely starred list of companies with non-whiteboard hiring processes), noodle (“Orchestrate agents using skills”), and how (“skill for explaining architecture”). [2]

What beginners can learn from her approach

(eTarget’s analysis, not a statement from Lauren Tan.) The recurring theme in her public talks is verification: give AI clear constraints and checks rather than trusting its output. That applies at every level. Even if you never write code, asking AI to mark missing facts as [NEED PROOF] and checking its sources is the same principle. See our Prompt Library and Grok vs Grok Bot.

TODO for the owner

• TODO: Confirm the “SpaceXAI” naming and her current title directly from her profile before publishing.
• TODO: Consider asking her permission or linking her own preferred bio.
• TODO: Re-check the talk and summary sources [4][5]; the summary is third-party, not written by her.

Sources

1. Lauren Tan, LinkedIn profile: https://linkedin.com/in/laurenelizabethtan
2. GitHub profile @poteto: https://github.com/poteto
3. LinkedIn post, Aug 2026: https://www.linkedin.com/posts/laurenelizabethtan_how-cursor-turned-ai-agents-into-better-engineers-activity-7493165544628703233-ZfM-
4. Video, “Lauren Tan XAi Grokbot”: https://www.youtube.com/watch?v=Cmoh-yR-usA
5. XiaoHu, “How a 24/7 agent team takes shape”: https://best.xiaohu.ai/en/article/cursor-agent-team-24x7/
6. X account @poteto: https://x.com/poteto

Sources checked October 12, 2026.