【Collaboration Request】Project Collaboration Request from Columbia University

Hi AIBrix team,

My name is Shen Li, an MSCS student at Columbia University. Our team of 4 is working on a final project for the High Performance Machine Learning (HPML) course, focused on:

**Performance optimization of LLM inference on accelerator-enabled infrastructure.**

We found AIBrix highly aligned with our interests, especially around scalable inference systems (routing, KVCache offloading, benchmarking, heterogeneous GPU support).

This project is supervised by [**Dr. Kaoutar El Maghraoui**](https://www.linkedin.com/in/kaoutar-el-maghraoui/) (Adjunct Professor of Computer Science at Columbia University and Principal Research Scientist at IBM Research).

We would love to contribute through a project-sized issue assigned by the team, with a lightweight mentor check-in (~30 min weekly or bi-weekly).
We will handle implementation, profiling, benchmarking, and deliver a PR + final report.

Preferred topics include KV cache–related optimizations, distributed inference, and LLM gateway & routing.

Our project timeline runs from **Feb 16 to May 4**, so we’re hoping to start soon.

I also previously worked 5 years at TikTok as a backend engineer, and we’re very excited to collaborate with the AIBrix community.

If there is a suitable issue or area to work on, we can start immediately.

Best regards,
Shen Li
MSCS @ Columbia University
Email: sl6008@columbia.edu
Linkedin: https://www.linkedin.com/in/shen-li-83714a384/

[HPML_Mentor_Guide.pdf](https://github.com/user-attachments/files/25218095/HPML_Mentor_Guide.pdf)