Haeshitha Indukuri

About

Built and productionized a MediCloud/Medicoud LLM microservice platform that lets clinicians query medical data in natural language, orchestrating multi-step RAG-style workflows with LangChain and evaluating/debugging with LangSmith. Delivered measurable gains (consistency ~70%→90% / +20%; latency ~2.0s→1.1s / -40%) by implementing structured prompts, fallback logic across multiple LLMs, hybrid retrieval tuning, and AWS Lambda performance optimizations (package size, async, caching).

Experience

Education

Key Strengths

• Built and deployed LLM-based medical data query microservice platform on AWS Lambda
• Improved response consistency by ~20% via prompt standardization and fallback logic with structured outputs
• Reduced latency by ~40% through Lambda optimization, async tasks, and caching
• Designed modular multi-step LLM workflows (retrieval + prompt construction) using LangChain
• Used LangSmith for tracing, debugging chains, and structured evaluation to find workflow inconsistencies
• Data-driven model/retrieval/prompt selection via small experiments balancing accuracy, latency, and cost
• Cross-functional collaboration translating clinician needs into prompts, retrieval tuning, and evaluation criteria

Reference Highlights

• efficiently improved PhD students' research work
• works independently with strong initiative
• strong project structure/design for LLM systems
• compassionate teaching style
• top-notch communication skills
• highly involved/engaged in assigned work
• able to make complex topics easy to understand using everyday examples
• collaborates effectively with people at any level