Intern in the field of GraphRAG-based AI Assistants

What you'll work on

• Analyze the existing GraphRAG (Retrieval-Augmented Generation) architecture to identify latency bottlenecks, specifically focusing on the efficiency and frequency of local LLM calls.
• Research state-of-the-art LLM optimization techniques, including model selection and quantization, prompt chaining, and local instance acceleration.
• Design and implement a performance-enhanced LLM pipeline that significantly reduces inference time for real-time shop floor (assembly operations) applications.
• Develop automated testing pipelines to conduct comprehensive latency studies and benchmark the new concept against the legacy workflow.
• Optional: Conduct usability and evaluation studies to ensure the improved response times meet the practical needs of hands-free, worker-led assembly processes.

What we're looking for

Student