Adusumalli V N Sai Lakshmi Pavan Kalyan
Hyderabad
Summary
Results-driven AI engineer with over 4 years of software engineering experience specializing in agentic AI applications. Proven track record in developing production-ready AI systems utilizing LLMs, RAG, and AI agent orchestration. Strong programming skills in Python, Java, Spring Boot, React.js, SQL, and cloud technologies facilitate the delivery of scalable enterprise-grade AI solutions.
Overview
1
1
Certification
4
4
years of professional experience
Work History
Software Engineer I – Deloitte
Medicaid – State of Wyoming
Hyderabad
11.2025 - Current
- Built Supervisor-based multi-agent insurance claims processing system using LangGraph, orchestrating Document and Medical Agents to automate claim analysis, enhancing overall processing efficiency.
- Designed LangGraph workflows with conditional routing and shared state to enable agents to process claims through LLM-powered analysis, improving claim handling accuracy.
- Implemented a Medical Agent with RAG using FAISS and embeddings to retrieve relevant medical knowledge and evaluate whether diagnosis, treatment, and supporting medical evidence are consistent.
- Developed a Document Agent with PDF parsing, OCR, and LLM-based structured extraction to extract patient, diagnosis, treatment, billing, and claim information from hospital bills, discharge summaries, prescriptions, and lab reports.
- Implemented structured LLM outputs using Pydantic schemas to produce consistent, machine-readable document assessments.
- Created Recommendation layer to consolidate outputs from multiple agents and generate claim recommendations with supporting reasons, facilitating informed decision-making.
- Planned the system for FastAPI-based integration and production deployment, with claim-level orchestration through a Supervisor Agent.
Systems Engineer – Tata Consultancy Services
Dicks Sporting Goods
Hyderabad
08.2022 - 11.2025
- Built an enterprise Retrieval-Augmented Generation (RAG) assistant supporting natural language search over technical documents.
- Designed document ingestion pipelines with recursive and parent-child chunking strategies to maintain context across multiple chunks.
- Implemented semantic retrieval using vector embeddings stored in Pinecone database.
- Combined vector search with BM25 retrieval, enabling hybrid search capabilities that leverage strengths of both methods.
- Implemented multi-level Redis caching for embeddings, retrieval/reranking results, and final LLM responses, reducing repeated-query latency from ~2–4s to
- Integrated query rewriting and reranking techniques, enhancing retrieval precision by refining search results.
- Applied prompt engineering techniques and optimized parameters such as temperature, Top-K, and Top-P to improve response quality, consistency, and reduce hallucinations.
- Utilized Langsmith to trace the application, create datasets and evaluate the accuracy of application.
Education
Bachelor of Technology - Computer Science And Engineering
Jawaharlal Nehru Technological University
Kakinada01-2022
Skills
- AI & LLM Frameworks: LangChain, LangGraph, LangSmith, MCP, Prompt Engineering, Observability, Evaluation, Structured Output, Tool Calling, OpenAI
- Retrieval & Search: RAG: Pinecone, PostgreSQL pgvector, Elasticsearch, FAISS, BM25, Hybrid Search, Query Rewriting, Cross-Encoder Reranking
- Backend: Python, Java, FastAPI, Flask, Spring Boot, Microservices, REST APIs, Redis
- Database: PostgreSQL, MySQL, MongoDB, SQL Server
- Cloud: AWS, AWS Bedrock, Docker, Git, GitHub Actions
- Tools: Postman, Github, Github Copilot, Github Desktop
Certification
- AWS AI Practitioner - AWS (Amazon Web Services)
- 2025, Microsoft Technology Associate – Python
Roles and Responsibilities
- Analyzed clients’ existing business processes to propose GenAI use cases and solutions that could improve efficiency and business outcomes.
- Developed agentic AI workflows and Retrieval-Augmented Generation (RAG) applications using Python, LangChain, LangGraph, with LangSmith for tracing, monitoring, and observability.
- Evaluated and validated GenAI applications using LangSmith to monitor performance, assess outputs against expected requirements, and identify areas for improvement.
- Collaborated within an Agile team environment, utilizing Git for version control to streamline development.
Timeline
Software Engineer I – Deloitte
Medicaid – State of Wyoming
11.2025 - Current
Systems Engineer – Tata Consultancy Services
Dicks Sporting Goods
08.2022 - 11.2025
Bachelor of Technology - Computer Science And Engineering
Jawaharlal Nehru Technological University