Souvik Kundu · AI PM Portfolio · Product 3 github.com/souvikkai
◆ RAG + LLM Benchmark
Resume × JD Matching Benchmarked Across 3 Configs
Paste a job description. The system retrieves your most relevant experience and generates tailored talking points — then benchmarks Claude Sonnet against Llama 3.1 8B on Groq for cost, latency, and quality.