My name is YuHang Wang (Ayanami1314), an undergraduate student at Shanghai Jiao Tong University (SJTU) majoring in Software Engineering (SJTU-SE).
I'm passionate about Large Language Models (pre-training, post-training, RL) and AI for Software Engineering (code agents, software engineering automation).
Currently, I'm a pretrain data researcher at Xiaohongshu Dots, working on pretrain data (code) and midtrain. Previously, I interned at Xiaomi MiMo, ByteDance (Douyin), Metabit Trading, and BondiTech.
I'm a strong believer in open-source culture. Check out my selected projects below!
I'm writing blogs on my website — notes and ideas from learning and coding.
| Area | Topics |
|---|---|
| LLM | Pre-training, Post-training, RL |
| AI for SE | Code Agents, Software Engineering Automation |
| Systems | Operating Systems, Agent Native Infra |
Also, looking forwards to some AI for Game cooperations(As a game enthusiast).
| Project | Role | Description |
|---|---|---|
| SWE-Pruner | First Author | Context pruning tool for LLM-based software engineering tasks |
| SWE-Pruner Pro | First Author | LLM-native context pruning for tool calling |
| SWE-Explore Bench | Co-First Author | Benchmarking how coding agents explore repositories |
| FastContext | Contributor | Training efficient repository explorer for coding agents |
(So why Ayanami? ayanami_(azur_lane) is my favorite anime character ^^. Just look my avatar.)