Dushyant

• Designed a retrieval-augmented system using Applied ML, Gen AI and LLMOps by integrating vector databases (e.g., FAISS) with LLMs for domain-specific question answering.

• Explored open-source LLMs (LLaMA / Mistral) and compared them with API-based models for cost-efficiency and scalability.

• Built and refined prompt-engineering pipelines to reduce hallucinations and improve contextual accuracy in AI outputs.