Tianle Li

Tianle Li

Member of Technical Staff @ xAI

firstlast@berkeley.edu

About

I am a Member of Technical Staff at xAI, currently on the Science of RL team. I also work on post-training and reasoning.

My contributions since May 2025:

  • Grok 4.2 Co-creator, only person running large-scale expert merging.
  • Grok 4.1 Fast Co-creator, led post-training RL training and recipe studies.
  • Grok 4 Fast Co-creator, led post-training RL training and recipe studies, co-led distillation and evals.
  • Grok 4 Core contributor, synthetic datasets, evals, tool use.

Background

Previously, I was an EECS undergraduate at UC Berkeley (2021-2025), where I was fortunate to be advised by Ion Stoica and built LMArena.

During my undergrad, I also spent a year full-time at Nexusflow as part of the LLM post-training team, working with Banghua Zhu and Jiantao Jiao. Additionally, I briefly interned as a student researcher at Google AI Research, where I was working on reasoning.

Research Interests

I am very interested in fundamental problems in training large models and solving superintelligence.

Some problems I've been thinking about:

  1. Can get more than 1-bit of information per rollout in RL?
  2. Can we skip mid-training? Can model learn on the fly?