(robot = continual+generative+reinforcement) * learning

Alan Chen

I teach robots to build things.

Undergrad researcher at UT Austin (Turing Scholars), working on RL for vision-language-action models at the Learning Agents Research Group. Previously at CMU's Robotics Institute, where I published on robotic assembly before finishing high school.

Research About me

currently

Learning Agents Research Group · UT Austin

RL for vision-language-action models

Fine-tuning robot foundation models (Octo, OpenVLA) with PPO, GRPO, and Diffusion-PPO for mobile manipulation in ManiSkill.

Arm · Central Engineering AI · Summer 2026

LLM Research Engineering Intern

Just wrapped up: independently developed a recursive language model approach and analyzed complex workloads to optimize uArch.

selected research

Robots that build things

Deep RL for robots that plan and build in the physical world: assembly planning, learning from demonstration, and policies that respect physics.

Planned sequences executed on the lab arms — vessel, guitar, faucet, branched stairs (16× speed)

RA-L · IROS 2025

Physics-Aware Combinatorial Assembly Sequence Planning using Data-free Action Masking

In what order should a robot place Lego-like primitives so the build never collapses halfway? We train a deep-RL construction policy guided by an online, data-free action mask that filters physically invalid moves before they happen.

AssemblyComplete: 3D Combinatorial Construction with Deep Reinforcement Learning: incomplete — The completion pipeline: library matching → scaled registration → constrained PPO construction

MECC 2025

AssemblyComplete: 3D Combinatorial Construction with Deep Reinforcement Learning

Hand a robot a half-built chair, and it should figure out what you meant to build, then finish it.

The learned policy assembling a structure on the real FANUC arm (8x)

arXiv 2023

Simulation-Aided Learning from Demonstration for Robotic LEGO Construction

A human demonstrates a LEGO build; the robot learns the manipulation policy safely in simulation, then reproduces the build in the real world.

All research →

writing

Notes & writeups

Jul 27, 2026 · 1 min

Rebuilding this site (and how writing works here now)

This site is a pile of Markdown files now. The old version was a React app, and adding anything meant editing components. By the time I finished fiddling with JSX, I didn’t feel like writing anymor…

site

Jul 20, 2026 · 3 min

Move 37, Move 78

My new revamped website’s homepage animation replays the opening of Game 4 of Lee Sedol vs. AlphaGo and stops at Move 78 which is arguably one of the most famous moves played by a human, a wedge th…

gorl

note · updated Jul 2026 · 6 min

Probability, likelihood, and measure

Why probability and likelihood are different objects, what MLE is actually doing, and the measure-theory vocabulary that makes densities, PMFs, and Dirac deltas one thing.

math

All writing →

beyond the lab

Physical Health Playing basketball with grad students, learning how to fence, and lifting Gaming lol Former Top 200 Hypixel Bridge Duels, Teamfight Tactics Diamond (just started), League Of Legends Hard Stuck Plat, Brawl Stars Masters Rank Stocks + Venture Capital Part of UT's Texas Venture Group and just personally enjoy staying informed on news Just enjoy learning more :) Robot learning, efficient AI (MLSys), diffusion, etc.

Say hi

Always happy to talk robots, RL, Go, or whatever.

alanjiach@gmail.com @AlanJiaqiChen