Featured project
Smart Spend Dashboard
Impact: Reduces manual financial analysis from ~2 hours to ~2 minutes.
Processes statements entirely in-browser — zero backend data exposure.
- Built a privacy-first credit card analytics dashboard using Claude AI and Node.js that analyzes spending trends and categorizes transactions from uploaded financial statements.
- All parsing and AI inference happens client-side, meaning no personal financial data ever touches a server.
- Designed an interactive breakdown view with category filters, monthly comparisons, and actionable spend insights.
Node.js
Claude AI
JavaScript
Privacy-first architecture
Financial data parsing
View on GitHub
Featured project
Bug Report Triage Agent
Impact: Reduces manual bug report triage analysis from ~1 hour to ~1 minute — zero infrastructure to maintain.
- Built and deployed a Bug Report Triage Agent as a GitHub Action that automatically classifies severity, applies labels, and detects duplicates on every new issue — eliminating repetitive manual triage work.
- Cut time-to-first response from 1 hour to under 1 minute, with the agent running fully autonomously at ~$0.005 per issue.
- Deployed with zero infrastructure to maintain by leveraging GitHub Actions as serverless compute, triggered directly on issue events.
Python
GitHub Actions
Pydantic
Anthropic SDK
HTTPX
PyYAML
pytest
Claude LLM Model
Ubuntu
GitHub Token
View on GitHub
Technical skills
Python
Golang
Java
JavaScript
REST APIs
Postman
SQL
Node.js
Incident response
APIs
Jira
Linux / Bash
Git