Pin (Campion) Qian

Currently, I am a Research Engineer at Meta, working on in-house post-training as a service in MSL Infra, with a focus on agentic RL support. I am interested in how agents learn from interaction and reliably complete long-horizon tasks.

Previously, I worked on synthetic data at Meta and game AI at Tencent. I received my M.S. in AI Engineering from Carnegie Mellon University.

Education

Carnegie Mellon University Aug 2023 - Dec 2024

M.S. in AI Engineering, Electrical and Computer Engineering

University of Liverpool Sep 2019 - Jun 2023

B.S. in Computer Science

Industry Experience

Meta Feb 2025 - Present

Research Engineer, MSL Infra

Work on in-house post-training as a service, with a focus on agentic RL support.

Machine Learning Engineer, Core Ads Growth

Work on post-training for computer-use agents, spanning SFT and RL.

Machine Learning Engineer Intern, Core Ads Growth May 2024 - Aug 2024

Worked on synthetic data for LLM post-training.

Tencent Jul 2022 - Oct 2022

Machine Learning Engineer Intern, Game AI Research Center

Worked on PUBG game AI agents and action-model fine-tuning.

Selected Work

AgenticVBench-Omni: Evaluating Omnimodal Agents on Economically Valuable Work

ICLR 2027 submission, under review.

Evaluating whether omnimodal agents can complete long-horizon professional tasks.

Worked on agent evaluation and failure analysis in Harbor.

From deterministic to stochastic: an interpretable stochastic model-free reinforcement learning framework for portfolio optimization

Applied Intelligence, 2023.

Zitao Song, Yining Wang, Pin Qian, Sifan Song, Frans Coenen, Zhengyong Jiang, Jionglong Su.

Risk-aware stochastic policies for portfolio optimization.

Worked on environment and reward design, and stochastic policy evaluation.

Academic Service

Reviewer (2026): NeurIPS Workshops DiffuLM, PALM, and VLM4RWD; ICML Workshops DL4C and AI4Math; EMNLP Workshops REALM and GroundLM.