Bhavya Jethwa

About Bhavya Jethwa

I’m an AI Engineer focused on building real-world, production-ready AI systems that solve…

Experience

1. Associate Software Engineer

   IBM

   Jun 2024 — Present

   Built a RAG chatbot using Python, FastAPI, LangChain, Pinecone, PostgreSQL, and OpenAI, enabling 200+ users to query internal documents, accelerating decision-making.• Implemented RBAC and authentication (JWT & OAuth), ensuring secure access and preventing unauthorized attempts.• Enhanced document retrieval via a hybrid search pipeline (BM25 + vector re-ranking), improving query accuracy by 40%, and optimized context handling to boost response relevancy by 30%, reducing redundant queries and saving 20% annually in LLM/API costs.• Developed an Agentic AI Operations Agent with LangGraph and Bash, automating tasks like cache clearing and service restarts, cutting manual workload by 40% and P3 incidents by 45%, and automating impact analysis to reduce processing time by 90% while preventing 50–70% of missed dependencies.• Reduced MTTD and MTTR by 35%, increased MTTF, and achieved 99.2% uptime, while integrating LangSmith observability tools to trace LLM pipelines, debug faster, and lower API/LLM costs by 50%.• Built CI/CD pipelines with GitHub Actions, Docker, and AWS ECS/ECR, leveraging AWS services (ALB, CloudWatch, Secrets Manager, EC2, Fargate) to ensure high availability, scalability, and smooth deployments, improving release frequency by 25%.