Amarnath Yadav

About Amarnath Yadav

As a Senior Software Engineer at agilon health, I contribute to the development of high-throughput data ingestion pipelines leveraging technologies such as Python, Snowflake, and SQS. My role also involves designing automated ELT pipelines using Snowflake, Airflow, and DBT to support analytics teams, as well as implementing Kafka-Airflow integration for real-time data flow. Collaborating across teams, I focus on optimizing performance in streaming pipelines and deploying services in Linux environments using Docker and Bash scripts. A graduate of Inderprastha Engineering College, Ghaziabad, I bring expertise in systems design, problem-solving, and PL/SQL to my work. My approach is rooted in leveraging cutting-edge technologies to deliver scalable and efficient solutions. I am motivated to drive innovation and support teams in achieving operational excellence.

Experience

1. Senior Software Engineer Java Spring Boot 3 Ai Llm Microservices

   agilon health

   Apr 2023 — Present · Gurugram, IN

   Working on backend systems that power clinical decision-making and patient care workflows at scale. My focus is on distributed systems reliability, AI integration, and JVM performance.Key contributions- Resolved critical JVM heap memory issues by profiling with Java Flight Recorder, migrating G1GC to ZGC, and eliminating memory leaks - achieving 99% improvement in heap utilization and 80% reduction in GC pause time- Architected an AI-powered clinical insights engine using Spring AI and OpenAI GPT-4, enabling intelligent extraction of risk signals from unstructured EHR notes and reducing manual review time by 60%- Built a Retrieval-Augmented Generation (RAG) pipeline using LangChain4j and pgvector, allowing clinicians to semantically query medical knowledge bases and improving decision support accuracy by 40%- Engineered RESTful and event-driven APIs with Spring Boot 3 and Apache Kafka, achieving 20% faster response times via async processing and query optimization- Solved dual-write consistency using the Transactional Outbox Pattern, guaranteeing 100% message delivery between PostgreSQL and Kafka with zero data loss- Implemented Saga Pattern for distributed transactions across 8 microservices, ensuring eventual consistency with automated compensation and zero manual intervention during failures- Deployed Redis caching with cache-aside strategy, achieving 25% latency reduction on high-frequency endpoints- Built Docker + Kubernetes pipelines with HPA and rolling updates, cutting deployment time by 30% and achieving 99.99% uptime- Led system design reviews (HLD/LLD), architecture discussions, and production readiness evaluations- Mentored 3 junior engineers through code reviews, design workshops, and pair programming.Tech: Java 17+, Spring Boot 3, Spring AI, LangChain4j, OpenAI API, pgvector, Apache Kafka, PostgreSQL, Redis, Docker, Kubernetes, AWS