Agentic / Generative AI Engineer at LangChain, Inc.

Engineered production agent workflows with LangChain and function/tool calling, added retries, timeouts, and sandboxed execution to meet strict SLOs. Built multi-stage RAG across docs, code, and tickets, reducing hallucinations and support load. Implemented LangGraph for stateful loops and memory with human-in-the-loop approvals, including regression suites and rollback paths for safe releases. Standardized tool schemas and permissions via Model Context Protocol servers for Git, Slack, and internal tools. Led migrations from Assistants API to Responses API while preserving contracts and memory semantics. Established rigorous evaluation and observability with LangSmith and custom harnesses, including golden datasets and cost/latency dashboards. Deployed agent APIs on AWS and GCP with Terraform, Helm, and GitHub Actions. Integrated multiple model providers behind a unified router with per-tenant keys and region-aware failover. Published open-source adapters for vector stores and tooling

Senior AI Consultant at Accenture

Led end-to-end delivery of multiple production chatbots from PoC to full-scale deployment and ongoing operations. Built a RAG-based medical chatbot ingesting patient records and images, performing symptom analysis, adaptive questioning, and candidate diagnoses through retrieval + LLM reasoning. Implemented core GenAI engine with OpenAI, LangChain, and Pinecone, including embeddings, retrieval orchestration, and grounding. Applied prompt engineering and function calling to enforce structured workflows and safer responses. Delivered features like next-prompt suggestions, per-user chat history persistence, and automated chat summaries. Developed a customer support & sales chatbot connected to a knowledge base, enabling vehicle recommendations. Enhanced numeric/mathematical query accuracy by translating NL to SQL via OpenAI and implemented semantic routing for pre-sales flow. Expanded omni-channel experiences (SMS/voice), contributing to increased sales.

Machine Learning Engineer at Google DeepMind

Developed internal AI chatbots to support operations and employee workflows, improving efficiency and response times. Built data workflows for collection, cleaning, labeling, and dataset versioning. Applied prompt engineering and fine-tuning with controlled experiments to optimize quality, latency, and cost. Implemented MLOps on AWS with CDK, SageMaker, Lambda, and EC2 for hosting, batch processing, and event-driven integrations. Created benchmarking dashboards to monitor model and system performance and integrated chatbot experiences into internal applications.