Pranav Paharia@pranavpaharia

Setting up AI and ML environments shouldn't be a bottleneck. Yet, developers frequently struggle navigating fragmented documentation, complex compatibility matrices, and architectural differences between AMD Instinct™ and Radeon™ RX GPUs. ROCm-Pilot eliminates this friction as an intelligent, context-aware setup assistant built exclusively for the AMD ecosystem. ROCm-Pilot is a Retrieval-Augmented Generation (RAG) agent that automatically and safely introspects local hardware. Using read-only diagnostic tools like rocm-smi and rocminfo, it detects your exact GPU architecture (e.g., gfx942), OS, and drivers. Rather than hallucinating, the agent grounds its answers in an auto-updating vector database scraped from official AMD GitHub repositories and developer blogs. By utilizing a hybrid search pipeline—combining GPU-accelerated ChromaDB semantic embeddings with BM25 sparse keyword matching—it ensures pinpoint accuracy for release-specific nuances. The platform outputs perfectly tailored, copy-pasteable bash commands and configuration scripts with verifiable citations. ROCm-Pilot also leverages Astral’s uv for blazing-fast dependency isolation. To ensure strict data privacy, ROCm-Pilot features a dual-tiered LLM architecture. It runs entirely locally by inferencing Google’s Gemma-4 (12B) natively on AMD hardware. For headless environments, it seamlessly falls back to cloud APIs via the Lemonade SDK and Fireworks AI. We also built a custom Evaluation Harness to continuously test RAG retrieval accuracy against a golden dataset of AMD developer queries. By automating complex configurations, ROCm-Pilot accelerates developer onboarding across the entire AMD AI ecosystem.

To participate in the hackathon, simply click Sign up with AMD. If you are not already a member of the AMD AI Developer Program (ADP), you will need to create an account. ADP members can access AI tutorials, experts, and community support. In addition, new members can claim a free month of DeepLearning.ai pro, $100 in AMD GPU cloud credits, and $50 in Fireworks AI API credits. All hackathon participants will receive on hackathon launch additional compute and API credits for their projects. Ready to build what's next on AMD? Sign up with AMD below to secure your spot.