Nishan S.

I build practical AI solutions that help businesses save time, automate work, and make better decisions using their data. My focus is not generic “chatbots”. I build AI systems that can actually understand your documents, answer questions accurately, extract structured data, and automate repetitive processes—so your team gets real results. I’ve worked across LLM apps, RAG pipelines, document understanding, and computer vision, and I’m comfortable taking a project from idea to a clean working solution with clear communication and regular updates. ⭐ What I Build • RAG Chatbots / Document Q&A • LLM Agents & Automation • Document AI & OCR • Computer Vision Solutions • Custom AI Integrations ⭐ AI Skills • Comfyui • txt-vid, img-vid, vid-vid • audio to text and text to audio • Prompt engineering + structured JSON output • Embeddings + vector search + retrieval tuning for accurate answers • LangChain / agent orchestration and multi-step reasoning flows • OCR pipelines and document parsing logic for messy real-world files • Computer vision ⭐ Tools & Tech Python, FastAPI, PyTorch, TensorFlow, OpenCV, YOLO, LangChain, LangGraph, FAISS, PostgreSQL/PGVector, MongoDB, n8n, Zapier-style automation ⭐ How I Work • Understand your goal and success criteria first • Start with a small working version quickly • Improve accuracy, speed, and reliability step by step • Keep everything simple, readable, and easy to maintain If you tell me what you’re trying to build (chatbot, OCR extractor, automation agent, vision system) and your data source (PDFs, images, database, website), I’ll reply with a clear plan and the best approach.