Software Engineer, AI at DoorDash

Engineered an automated support resolution engine using LlamaIndex, LangChain, and Python, integrating a multi-agent RAG architecture over millions of vector embeddings to reduce Dasher issue resolution time by 28% and minimize manual escalations. Fine-tuned enterprise LLMs with PyTorch and LoRA to extract entity schemas from 500,000+ retail SKUs and build a scalable product knowledge graph on AWS EKS, improving cross-vertical search relevance by 14% for grocery rollouts. Architected an automated LLM simulator using CrewAI agents to generate synthetic multi-turn conversations, validating chatbot workflows via Docker and Kubernetes across complex hallucination vectors. Optimized a logistics predictive routing service in Go/gRPC to deliver real-time Dasher ETA inference, sustaining 1M+ QPS with p99 latency under 45ms across distributed AWS infrastructure. Built a low-latency feature aggregator decoupling events via Kafka and Redis and persisting engineered features into PostgreSQL for sy

Software Developer at Oracle

Developed a financial reconciliation dashboard using React.js, Java Spring MVC, and Oracle DB to process 500,000+ daily ledger entries and manage reconciliation exceptions efficiently. Designed and implemented scalable RESTful APIs with Spring Boot and Node.js to support secure data exchange between legacy Oracle databases, internal financial platforms, and cloud-ready microservices. Built interactive financial reporting interfaces with Angular and Bootstrap, using client-side performance optimizations to reduce rendering delays by ~20–25%. Implemented asynchronous background processing with Amazon SQS to decouple heavy reporting/reconciliation jobs, improving UI responsiveness and reducing primary web-server load by ~30–40%. Applied TDD practices using JUnit/Mockito and frontend test automation, improving test coverage to 85%+ and reducing regression risk in Agile/Scrum delivery via GitLab CI/Maven pipelines.