Yufeng He · 何宇峰

Building reliableagent systemsfrom post-training to inference.

I work at the intersection of agent algorithms and AI infrastructure. At Moonshot AI (Kimi), I worked on coding-agent post-training, self-improvement, and agent applications. As a Mooncake maintainer and contributor to vLLM and PyTorch, I connect model iteration with reliable execution.

Agent harnessesCoding-agent post-trainingMulti-agent systemsLLM systems
View résumé ↗
Emailheyufeng2003@qq.com
GitHubgithub.com/he-yufengLinkedInlinkedin.com/in/yufenghe
M.Sc. Computer Science · The University of Hong Kong · Expected Dec 2026
Yufeng He
Agent systemsOpen sourceHong Kong
560+upstream contributions
200+agent-framework & eval PRs
100+model-systems contributions
1.8K+CoreCoder stars

02 / Open source

Open source,
at systems depth.

560+ upstream contributions across 67 projects in the Agent and LLM ecosystem, from tool-call lifecycles to CUDA kernels and KV-cache transfers.

Maintained & contributed upstream

Contributions across
the Agent systems stack.

560+merged or landed PRs
across 67 projects
01

Frameworks & evals

200+

Merged upstream PRs

Qwen Code ↗60 PRsTool-call state & cancellation
DeerFlow ↗25 PRsConcurrency & checkpoints
Microsoft Agent Framework ↗35 PRsAgent-loop boundaries
OpenClaw ↗9 PRsSession safety & permissions
02

Model systems

100+

Code, tests & tooling · Mooncake OSS Maintainer

Mooncake ↗59 PRs · MaintainerDistributed KV-cache storage & transfer
vLLM ↗17 PRsRequest isolation & CUDA Graph stability
PyTorch ↗16 landed PRsCUDA kernels & Inductor correctness
Triton ↗8 PRsGPU compiler semantics
03

Agent applications

100+

Merged upstream PRs

Vibe-Trading ↗71 PRs · MaintainerPortfolio analytics & trading gates
AstrBot ↗37 PRsProvider & scheduler reliability

Merged PRs and bot-landed changes; selected projects shown · Full contribution record ↗ ·

Independent open source01

Creator & Maintainer

CoreCoder ↗

Independently designed and built a compact Python coding-agent engine with task planning, tiered context management, and multi-agent collaboration. A runnable harness for understanding and extending the agent loop.

1.8K+GitHub stars

Python · Runnable agent harness

Core development & maintenance02

Core Developer & Maintainer

Vibe-Trading ↗

A multi-agent platform for automated investment research and quantitative trading, connecting financial analysis, strategy research, and trading workflows.

37K+GitHub stars

71 merged PRs · Agentic finance

01 / Experience

Research depth.
Engineering range.

Across agent harnesses, coding-agent post-training, multi-agent systems, agentic finance, and LLM infrastructure.

Kimi

Moonshot AI

AI Agent Research Intern

Agent research that survives contact with engineering.

Coding Agent Algorithms Team

Base-model post-training

  • Independently built post-training datasets and benchmarks for long-horizon coding tasks, automating data generation and evaluation.
  • Explored self-improving agent post-training; cross-evaluated models and harnesses to validate and refine training recipes.

Alpha Lab

Agent algorithms, harnesses & applications

  • Core contributor to finance agents for automated investment research, factor discovery, and portfolio optimization, from algorithm validation to system implementation.
  • Independently led multi-agent and agent-swarm experiments, exploring group coordination and evolutionary optimization.
01

Baidu

AI Agent Engineer
AI Talent Program

Co-developed an advertising bidding agent that jointly modeled ROI and quality scores, improving the team's business ROI target attainment from 81% to 101%.

02

Maimai

AI Agent Engineer
CEO Mentorship

Under CEO mentorship, independently built a skill-representation agent and quality-evaluation pipeline, improving the automated career-recommendation loop.

03

Kuaishou

Machine Learning Engineer
Transcoding scheduling

Contributed to a popularity-aware scheduling agent for hundreds of thousands of transcoding CPU cores, jointly modeling compute costs and bandwidth benefits.

05 / Research

Research with
testable claims.

My work connects coding-agent evaluation, collective learning in multi-agent systems, and AI-assisted mathematical discovery—building experiments and verifiable artifacts around each research question.

R/01

First author

TrajBias ↗

Built a coding-agent trajectory evaluation framework, identified scoring biases, and evaluated debiasing strategies through controlled comparisons of task trajectories.

Submitted · ACL Rolling Review, October 2026

R/02

Project lead

AgentSociology

Built a social-evolution testbed with agent collectives as the unit of study. Exploring how autonomous interaction and experience transmission shape self-organization, collective learning, continual evolution, and emergent intelligence.

Multi-agent systems · Group-level evolution

R/03

Project lead

AlphaFormal

Exploring a decentralized AI mathematical-discovery network where verifiable proofs establish trust between agents; connecting conjecture exploration, counterexample search, and proof reuse.

Three first-author manuscripts submitted to SCI-indexed mathematics journals · Lean checks of selected claims

R/04

Sole author

FMGAN ↗

Adapted adversarial refinement to time-series imputation model training, connecting adversarial learning with reconstruction of missing observations.

KDD 2026 MiLeTS Workshop

04 / Selected projects

Built beyond
the benchmark.

Two hands-on projects in health screening and AI music generation, combining model work with a usable application.

CTO · Jun 2025 — Jul 202601

Edge AI · Health screening

FormulaSperm

Led an edge-AI health self-screening system from scratch, spanning dataset construction, model training, and on-device inference optimization. The project won HKU DeepTech100.

Co-founder · Full-Stack Engineer
Oct 2021 — Jun 202202

Generative AI · Music systems

Qiyin Tech

Co-founded an AI music-generation project and built its generation engine and application. The project won National Gold at China's “Internet+” competition.

03 / Education

Sep 2025 — Dec 2026 (expected)

The University of Hong Kong

M.Sc. in Computer Science

Sep 2021 — Jun 2025

Beijing University of Posts and Telecommunications

B.Eng. in Telecommunications Engineering

Selected honors

  • Champion2026 HKU DeepTech100 · HK$1M+ prize
  • Captain · 1st Place (Gold)BUPT · 2023 Shanghai Global AI Innovation Competition · LLM Track
  • 3× SilverACM ICPC Asia Regional Contests
  • National First PrizeNational Olympiad in Informatics in Provinces (NOIP)

03B / Skills & interests

Fluent across the stack.

Languages

Python / C++ (primary) · Java / Go

Agent systems

Agent harnesses · Coding-agent post-training · LLM systems

Practice

Claude Code / Codex power user · English (professional working proficiency)

Interest

National Top 10 · China Children's Go (Weiqi) Grand Prix