Experience

Software Engineer 1 · New York, NY

• Full-time role after graduating from UMD.

Software Engineering Intern · Pleasanton, CA

• I owned the whole monitoring stack — ingestion, storage, AI diagnostics, alerting, dashboard. Nobody else was building this; it was my thing from day one.
• Built a chatbot on AWS Bedrock + Claude that engineers actually used to debug cloud issues. The AI part was straightforward — the hard part was making the knowledge base good enough that the answers weren't garbage.
• Wrote a pipeline with Lambda, S3, and CloudWatch that kept the chatbot's context up to date automatically. Before that, it went stale within a week and nobody wanted to update it manually.
• Got BHN from partial to 100% endpoint monitoring by deploying ThousandEyes agents across every region. Wrote a CloudFormation Stack Generator to do it without clicking through the console for each one.
• Built the React + Flask dashboard. Detection time went from hours to basically instant — engineers saw outages as they happened instead of hearing about it from customers.

Software Engineering Intern · San Francisco Bay Area, CA

• The best decision I made here wasn't technical — it was shipping our LLM tools inside Microsoft Teams instead of building a separate app. That's why it hit 10k+ daily queries. People used it because it was already where they worked.
• Built a GitHub PR reviewer in Python + Claude that analyzed repos and spit out code review feedback. It actually got used day-to-day, not just demoed.
• Plugged the LLM into Teams through Ciena's internal Navi chatbot API so the whole company had access without installing anything.

Software Engineering & Data Science Intern · San Francisco Bay Area, CA

• Built a ticket classifier with Claude and Langchain — 99% accuracy across 20+ categories. But the real win was that it kept working after I left. Automation that runs without you is worth way more than automation that needs babysitting.
• Built an FAQ system that handled common issues before they hit a human. Escalations dropped 50%, response time dropped 75%.
• Set up Databricks + MLflow for model versioning and batch processing so the whole thing could scale without someone manually managing it.

Lead Software Developer / Founder · San Francisco Bay Area, CA

• Started this from nothing while I was still in school. It watched retailers for price drops, restocks, and promo codes, then secured deals for customers automatically.
• The backend was Go with gRPC, custom HTTP clients, and proxy rotation — millions of requests a day. The hard part was doing that at scale without getting blocked, which meant TLS fingerprinting, smart proxy rotation, and connection pool management.
• 100k+ deals secured. I handled everything — code, infra, pricing, customer support. Learned more here than any classroom.

Software & Data Research Intern · Berkeley, CA

• First real engineering job. Processed 1TB+ of microscopy data with Python and NumPy for CNN training sets. Biggest lesson: be paranoid about data pipelines. One bad transformation and you've corrupted thousands of images that took weeks to produce.
• Built segmentation tools for labeling research images. The work directly supported published papers. Taught me to validate early, validate often.