Team Idea
Building a multimodal RAG system for Mini-Challenge 3 on AMD ROCm. We already have a working end-to-end pipeline: a crash-proof parser for mixed documents (PDF, DOCX, XLSX, CSV, logs, code), a vision-language model for images, and a resident model server that stays loaded between questions. We're looking for someone with hands-on AMD/ROCm GPU access (MI300X or Navi 31) for build and VRAM testing, and someone strong in retrieval or LLM evaluation to stress-test answers against harder corpora. Async-friendly; we split work by component.
Team Members
Harsh Bajaj
harsh_p_bajaj258
Submission
Team Leader hasn't made a submission yet