Headline

Lead AI Data Scientist & Solutions Architect @ FarmwiseAI

About

As a Lead AI Data Scientist & Solutions Architect at FarmwiseAI, I design and deliver production-ready AI products, leveraging expertise in SaaS development, solution architecture, and PostGIS. With multiple Azure certifications, I contribute to cutting-edge projects across geospatial, legal, and video AI, driving cost optimization and operational efficiencies. My work includes reducing LLM inference costs by 45% and standardizing internal architectures to streamline new use-case integrations. Collaborating with enterprise and government clients, I translate complex, ambiguous challenges into scalable, deployed solutions. By fostering teamwork and innovation, I aim to empower organizations with impactful AI-driven tools while championing reusable frameworks and efficient workflows. My goal is to continue building transformative technologies that align with organizational goals and make data-driven decision-making accessible and effective.

Location

Chennai

AI stack

gemini

langchain

langgraph

mcp

Work experience

Lead AI Data Scientist & Solutions Architect at FarmwiseAI, Mar 2025 - Present · 1 yr 4 mos, Chennai, Architect and lead a team building production AI products end to end — owning design, delivery, and stakeholder comms. - Shipped 8+ AI products in 12 months as both lead and IC across geospatial, legal, and video AI; cut time-to-prototype from weeks to under 5 days with reusable RAG and agentic templates. - Cut LLM inference cost ~45% through tiered model routing (open-source for low-stakes, frontier for production reasoning) and prompt-caching. - Standardized an MCP-based architecture letting any internal agent plug into shared tools — reducing new-use-case integration effort ~60%. - Own delivery directly for IAS-level government and enterprise clients, translating ambiguous problems into deployed systems in 4–8 week cycles. - Mentor 4+ data scientists on RAG design, LLM evaluation, prompt engineering, and full-stack deployment.

Lead Data Scientist at FarmwiseAI, 2025 - Present · 1 yr 6 mos, Role: Architect & lead a team building production AI products end-to-end; own design, delivery & stakeholder comms. Shipped 8+ AI products in 12 months as lead + IC across geospatial, legal & video-AI; cut time-to-prototype from weeks to <5 days via reusable RAG/agentic templates. Cut LLM inference cost ~45% through tiered model routing (open-source for low-stakes, frontier for production reasoning) and prompt-caching. Standardized MCP-based architecture enabling any internal agent to plug into shared tools — reducing new-use-case integration effort by ~60%. Mentor 4+ data scientists on RAG design, LLM eval, prompt engineering & full-stack deployment. Direct IAS / enterprise-client ownership — translate ambiguous problems into deployed systems in 4–8 week cycles.

Associate Data Scientist (AI/ML Engineer) at FarmwiseAI, Feb 2023 - Mar 2025 · 2 yrs 2 mos, Chennai, Built end-to-end AI systems as an IC — from problem framing through production deployment on AWS and GCP. - Delivered 12+ AI projects end to end — RAG, agentic, computer vision, OCR, and geospatial — owning the full lifecycle. - Built RAG over 100K+ documents with hybrid retrieval (FAISS + BM25 + re-ranking); achieved 90%+ citation accuracy on a legal corpus and saved ~70% of research time. - Designed VLM-based video analytics enabling natural-language queries over CCTV footage — cutting review from hours to under 30 seconds. - Productionized 20+ AWS Lambda + S3 services with cold-start optimization, sustaining 99.5%+ uptime. Pioneered Indic-language voice and chat pipelines (Sarvam STT + LLMs), expanding reach into Tier-2/3 markets.

Intern at Data Science and Analytics Centre, Apr 2022 - Mar 2023 · 1 yr, Coimbatore

Education

Bachelor of Technology, Artificial Intelligence, karpagam college of engineering