🏆 Notable contributions
Popular open-source repos they've shipped to (commits + PRs)
Score breakdown
🛠 Featured work
Their own popular and pinned repositories
Official implement of paper "Revisiting Multimodal Positional Encoding in Vision–Language Models", ICLR 2026
PythonSimple Implementation of Pix2seqV2(multi-task)
Python基于Netfilter的Linux状态检测防火墙，支持NAT。华中科技大学2023学年网络安全课程设计项目，参考https://github.com/RicheyJang/RJFireWall
Cπ₀ inference in C/C++, built on top of llama.cpp.
C++Self-aiming and perspective tools for Assault Cube ver1.3.0.2
C++dev
Python🧬 Stack & domains
🧬 Most similar developers
Closest profile, nearby score
🔥 Full roast
JJJYmmm — 79.10/100 · 人上人 (优质贡献者 · 值得信任)
TL;DR: Niche AI open-source contributor with a strong track record of substantive, well-accepted PRs to top-tier ML frameworks, no red flags, though original project star count is modest.
| Dimension | Score | Notes |
|---|---|---|
| Account maturity | 8.5/10 | Registered ~4.7 years, active across 6 contribution years, last activity 1 day ago |
| Original project quality | 10.4/18 | 10 non-empty original repos, 128 total stars, top repo Multimodal-RoPEs (ICLR 2026) with 88 stars |
| Contribution quality | 19.2/27 | 38 merged PRs out of 55 total, 16% rejection rate, only 4 external trivial PRs |
| Ecosystem / maintenance impact | 20/20 | 16 substantive PRs merged into ★161k+ repos (Hugging Face Transformers, vLLM, llama.cpp, mlx-vlm), focused on Qwen3.5 series support |
| Community influence | 6.1/8 | 118 followers, 11 following, healthy follower-to-following ratio with no follow-farming迹象 |
| Activity authenticity | 14.9/17 | 266 contributions in the last year, active across 4 activity types, consistent recent activity |
Red flags: None Manual adjustment: None (0), as contributions are substantive, no templated PR flooding or trivial PR farming detected, and all red flag triggers are unset. Verdict: Prioritize, trusted niche AI open-source contributor with proven expertise in multi-framework model integration.
🔥 Roast: You’ve spent the past year spamming well-crafted Qwen3.5 support PRs across every major ML framework like an AI gig worker chasing contribution KPI, and your 88-star ICLR 2026 repo is the only original work anyone’s ever bothered to star — at least you’re good enough at your day job that the maintainers keep merging your patches instead of marking them as noise.