Amit Seth
AI/ML & Prompt Engineer specializing in Agentic workflows, Retrieval-Augmented Generation (RAG), and LLM orchestration
Hire this person
Send a job offer directly to this candidate
About
Detail-oriented AI/ML & Prompt Engineer specializing in Agentic workflows, Retrieval-Augmented Generation (RAG), and LLM orchestration. Proven track record of architecting and deploying autonomous AI agents that solve complex enterprise problems, from data analytics to conversational sales. Active open-source contributor in the AI observability space.
Highly proficient in dynamically integrating foundation models (NVIDIA Nemotron, Gemini, OpenAI) into Python (FastAPI) and JavaScript (Node.js) architectures.
Experience
Freelance AI & Full-Stack Engineer - CodeNCraft (Self-Employed)
(2024-08)
- Enterprise RAG Copilot: Architected a privacy-first, end-to-end local knowledge assistant designed to ingest sensitive enterprise documents (e.g., Legal contracts, Hotel management manuals).
- Built a high-performance Python (FastAPI) backend integrating LangChain, Gemini Flash, and ChromaDB for secure, local vector embeddings and retrieval.
- Developed a responsive Next.js frontend with streaming Markdown rendering to provide a premium ChatGPT-like user experience.
- Autonomous Data Analyst Agent: Architected a production-ready AI agent capable of querying complex, private enterprise databases (SQLite) via natural language.
- Engineered a custom RAG pipeline that dynamically injects live database schemas into the LLM context window.
- Implemented dynamic model swapping, utilizing NVIDIA Nemotron-3 Ultra (550B) for complex relational logic and Gemini 2.5 Flash for high-speed, low-latency queries.
- Enterprise AI Sales Agent: Developed an autonomous, conversational sales widget for B2B client portfolios.
- Implemented negative prompting and rigid persona boundaries to prevent the AI from breaking character or answering off-topic queries.
Open Source Contributor - Traceloop OpenLLMetry
(2024-02 - 2024-06)
- Core SDK Optimization: Refactored the Python traceloop-sdk (a popular GenAI observability tool) to implement granular optional dependencies for 30+ AI and Vector DB integrations (OpenAI, LangChain, ChromaDB, etc.).
- Improved developer experience and significantly reduced package bloat by overhauling dynamic module loading logic and gracefully handling ModuleNotFoundError across the codebase.
Front-End Developer Intern - 1stop AI & Kapil IT
(2024-02 - 2024-06)
- Developed responsive, component-based user interfaces for web applications, laying a strong architectural foundation for later integrating complex AI/ML backend services into user-friendly SaaS platforms.
Education
Master of Science (M.Sc.) - Information Technology - Pt. Ravishankar Shukla University, Raipur (2022 - 2024)
Bachelor of Science (B.Sc.) - Computer Science - Pt. Ravishankar Shukla University, Raipur (2019 - 2022)
