RAG pipelines over 83,000 documents. ML that retrains itself from live feedback. Agents that debug their own code. Based in Phoenix, AZ, graduating May 2026.
MS Software Engineering · ASU · May 2026
I started with a B.Tech in Biotechnology, a discipline that trained me to think in systems, run controlled experiments, and never trust a result without stress-testing it. That instinct transferred directly into software engineering.
At 4 Systems Info Solutions I spent over two years building at real scale. Spring Boot APIs processing millions of records, database schemas tuned to cut query latency by 40%, and CI/CD pipelines that reduced release windows by 60%. Real systems, real traffic, real accountability.
Now at Arizona State University (GPA 4.0), I have gone deeper into intelligence: embedding 83,000+ FDA documents for RAG retrieval, architecting a 7-service ML pipeline that retrains from live user feedback, and writing a LangGraph agent that detects, diagnoses, and patches its own bugs.
I also taught 150+ students Java as an Instructional Assistant. The clearest proof you understand something is your ability to explain it clearly to someone who does not.
Mnemo, an AI assistant solving three production GenAI problems at once: persistent memory across sessions, RAG document grounding so it stays factual, and confidence-scored hallucination detection. Stack: LangGraph, Claude API, ChromaDB, PostgreSQL.
A personalized adaptive health chatbot built for industry sponsor MyEdMaster as part of a 6-person Agile team across 6 sprints. Real RAG pipeline. Real safety system. Real users.
An AI assistant solving three production GenAI problems simultaneously. Persistent memory across sessions so it actually remembers you. RAG document grounding so it stays factual and cites sources. Confidence-scored hallucination detection so you know when to trust it.
Self-improving ML recommendation system with 7 independent microservices in Docker Compose. User feedback streams via Redis Streams auto-trigger model retraining at 50+ events. RMSE 0.947 on 100k MovieLens ratings. Models hot-swap into production without downtime.
Autonomous LangGraph agent that detects failing tests, reasons through root cause, patches the code, and reruns CI without any human intervention. Infrastructure that maintains and repairs itself.
Accessible e-commerce platform in React 19 and TypeScript scoring 95/100 on Lighthouse, placing it in the top 2% globally. WCAG 2.1 AA compliant with voice command navigation, ARIA live regions, and colorblind modes built in from the start.