- Design, build, deploy and monitor AI applications end-to-end — chatbots, RAG-based assistants, document-intelligence pipelines, agentic workflows and internal tools.
- Write clean, tested, production-grade Python code and participate actively in code reviews.
- Build retrieval pipelines — chunking, embeddings, hybrid search, reranking — and iterate based on evaluation results.
- Integrate AI features with enterprise systems, APIs, databases and third-party platforms.
- Work with self-hosted open-weight models alongside API-based models; choose the right tool for the job based on cost, latency, privacy and quality.
- Instrument systems for latency (p50/p95), cost per request, memory and token usage, response quality and tool-call success rate — and act on what you measure.
- Build evaluation harnesses before shipping — not eyeball outputs.
- Contribute to fine-tuning and model-adaptation experiments under senior guidance.
- Sit with business users to understand real workflows and translate them into working prototypes quickly.
- Own on-call for what you ship; document your work so it can be handed off, extended and maintained.

📌 AI Engineer (Chennai)
🏢 Pon Pure Chemicals
📍 Chennai