Headline

AI/ML Engineer | Building production LLM systems & RAG pipelines | Fine-tuning with LoRA | GenAI + Computer Vision | Agentic AI

About

Most ML engineers use pre-built APIs. I fine-tune the models behind them. What I do I'm an AI/ML Engineer focused on building systems that actually work in production — not just Jupyter notebooks. My work spans three areas: fine-tuning large language models using LoRA (TinyLLaMA, Qwen) for real-world latency and cost constraints, building RAG pipelines that give LLMs reliable memory over large document stores, and deploying computer vision systems at scale — including 68-point facial landmark detection trained on 100K+ images. Who I help & how I've built NLP interfaces that let non-technical teams query databases in plain English, and optimized inference pipelines handling thousands of concurrent requests. My goal is always the same: reduce the gap between an impressive demo and a reliable, low-latency system someone can actually use. Stack & tools Python · PyTorch · HuggingFace · LangChain · RAG · LoRA / QLoRA · AWS (Certified AI Practitioner) · Computer Vision · NLP · FastAPI · MLOps

Location

Surat

Work experience

AI/Ml Engineer at eDelta, Nov 2025 - Present · 7 mos

AI/ML Intern at eDelta, Mar 2025 - Oct 2025 · 8 mos

Education

Master of science in AI & ML, Veer Narmad South Gujarat University, Surat

Bachelor of Computer application, Smt. Tanuben and Dr. Manubhai Trivedi College of Information Science

Public links

github: https://github.com/Harshvi25