Junjie Xiong
Profile photo of Junjie Xiong

I'm Junjie 俊杰.

CS + Maths @ UC Berkeley

My research centers on reinforcement learning, training language model agents, continual learning, and LLM evaluation.

More broadly, I am interested in language models and machine learning systems, and I am currently deepening my understanding of efficient inference.

Outside of research, I am drawn to investing and the venture capital world, and I am an avid reader. I also enjoy staying active through swimming, long-distance running, and racket sports like tennis, squash, and pickleball.

Currently

I am looking for roles in post-training research, machine learning engineering (MLE), and software engineering (SWE).

Research

Berkeley, CA Post-Training Researcher, Berkeley Sky Computing Lab Dec 2025 – Present

RL for language models and agents; advised by Raluca Ada Popa & Ion Stoica.
I am forever thankful for the mentoring from Sijun Tan, Tianhao Wu, and Sida Li.

Stanford, CA Student Researcher, Stanford AI Laboratory (SAIL) Jun 2026 – Present

NLP Group; advised by Dan Jurafsky and mentored by Chen Shani, exploring whether post-training can steer language models toward concept-preserving reasoning.

Florence, Italy Undergraduate Researcher, Duke University Jun 2024 – Aug 2024

Contributed to the Vulci project with Maurizio Forte, doing geophysical surveys and computer graphics; mentored by Nevio Danelon and Antonio Lopiano.

Internships

Palo Alto, CA Software Engineer, Tencent Jun 2026 – Present
San Francisco, CA Data Scientist, Epoch AI (Contract) Mar 2026 – Jun 2026

Built an evaluation harness for math reasoning that uses an automated Lean prover to auto-formalize natural-language math into Lean and machine-check the resulting proofs, enabling scalable, rigorous grading of model outputs.

Software Engineer, NVIDIA (Contract) Sep 2025 – Dec 2025

First-authored "Towards Expert Financial QA via Self-Improving RAG" (AFA Workshop @ ICLR 2026).

Beijing, China Software Engineer Intern, Xiaomi Technology Jun 2025 – Aug 2025

Two rotations:
1) AI Product, building a Plaud-style voice-memory product with a memory layer over Mem0, Zep, and Letta, and audio-language models for detection and evaluation;
2) AI Infra on Xiaomi MiMo, the company's large language model lab, building pretraining data pipelines (deduplication and quality filtering) and internal tooling for data QA, validation, and dataset access.

Shenzhen, China Summer Analyst, Cowin Capital Group May 2023 – Jul 2023

Supported deal sourcing and due diligence for early-stage consumer investments, including Tea'stone, Pop Mart, Chi Forest, and Qiandama. Cowin is a private equity and venture capital firm founded in 2000, with about US$4.7 billion in AUM.