I build software that works welland looks like someone cared.
Firmware validation tooling at Western Digital, and before that customer platforms in React and Next.js.
I like the backend to be solid and the front of it to feel good to use.
Chip Design Optimization
UCSD × Qualcomm
Research into training graph neural networks on large-scale chip design data, and a traversal algorithm for graphs too big to walk naively.
Read the algorithm write-upCustom batching strategy
Trained neural networks efficiently across 80GB of chip design data.
Weighted random walk
A statistical algorithm based on Markov chains that traverses 1M+ nodes.
Hardware design validation
Applied AI techniques to optimize the validation process itself.
Where I have worked
Three years shipping
production systems.
Western Digital
Software Engineer · Irvine, CA
- Contributed to an automated firmware validation framework in Python/Linux, increasing firmware stability for new enterprise storage drives.
- Engineered an autonomous RAG system for firmware and log analysis that classified over 10,000 failure patterns and cut root-cause triage time by more than 20%.
- Architected a standardized log-parsing pipeline leveraging MCP, integrating data into company AI agents.
Energy311
Software Engineer · San Diego, CA
- Built and deployed a production solar energy billing and monitoring platform with React and scalable Python APIs, supporting 2,000+ active customers.
- Architected secure integrations for payments, onboarding, and reporting, enabling compliant handling of sensitive transaction data.
- Migrated the company website to Next.js with the design team, achieving an 8% increase in active user sessions.
Inautics
Software Engineer · San Diego, CA
- Designed, programmed, and tested healthcare enterprise software processing more than 5,000 claims per day.
- Built a secure data pipeline ingesting healthcare data from a customer GraphQL API and SFTP server.
- Created an efficient EDI file parser in Rust that is HIPAA compliant and secure.
Firmware validation, payments infrastructure, healthcare claims.
What I work with
The stack behind
the systems.
Mostly Python and Rust against Linux systems, with TypeScript and React on the front end. Currently deepening the machine learning side at Georgia Tech.
Measured impact
Systems I have shipped classify 10,000+ failure patterns, cut triage time 20%, serve 2,000+ customers, and process 5,000 claims a day.
Firmware that ships more stable. Billing that reconciles. Claims that clear. The numbers come from production systems, not benchmarks.
Education
Computer science,
then data science.
M.S. Computer Science
Enrolled Aug 2024 · Remote
B.S. Data Science
Oct 2020 — Mar 2024 · San Diego, CA