Akshaia Sai H

AI/ML Engineer specializing in LLM application development, Agentic AI, and Retrieval-Augmented Generation (RAG) systems. Experienced in shipping production-grade backend services with FastAPI, LangGraph, and PostgreSQL, containerizing and deploying with Docker/AWS, and fine-tuning open-source LLMs (LoRA, QLoRA, SFT, DPO) for domain-specific tasks. Strong foundation in vector search, evaluation-driven development, and prompt engineering for agentic system design.

AI Developer Intern at Goodbooks ERP (2025-10 – 2025-11)

• Built an Agentic RAG chatbot (LangGraph) enabling natural-language Q&A over multi-page BOM PDFs via OCR extraction and retrieval-augmented generation, replacing manual document lookup.
• Fine-tuned local Qwen2.5-1.5B (Unsloth, SFT) to correct domain-specific factual errors, after evaluating and ruling out DPO as unsuitable for single-pair correction tasks.
• Evaluated retrieval quality via RAGAS (Context Precision, Faithfulness, Answer Relevancy) and iterated on chunking strategy and prompt design to improve answer accuracy.

M.Sc. in Artificial Intelligence & Machine Learning – Coimbatore Institute of Technology (2023 – present)