A selection of projects across systems engineering, developer tooling, and product infrastructure.
A fault-tolerant, horizontally-scalable job queue built for high-throughput workloads. Handles millions of tasks per day with guaranteed delivery, retry logic, and dead-letter queuing.mm
Designed the core scheduling algorithm and worker pool architecture. Reduced average job latency by 60% compared to the previous solution. Deployed across three AWS regions with zero-downtime rolling updates.
Live data visualization platform ingesting and rendering millions of events per day. Built for a SaaS product serving 50k+ active users.
Architected the WebSocket event pipeline and React rendering layer. Implemented a custom time-series aggregation engine that reduced query times from 4s to under 200ms. Designed the component library used across the product.
Lightweight, composable middleware layer for microservice orchestration. Handles auth, rate limiting, request routing, and observability in a single unified layer.
Built as an internal open-source project adopted by four product teams. Reduced boilerplate per service by ~70%. Includes a plugin system for custom middleware and a CLI for scaffolding new services.
Opinionated command-line tooling to streamline local development workflows — environment setup, service orchestration, and deployment pipelines in one place.
Used daily by a team of 20+ engineers. Reduced onboarding time for new developers from two days to under two hours. Includes interactive prompts, config validation, and integration with CI/CD pipelines.
Centralized logging, tracing, and alerting infrastructure for a multi-tenant SaaS platform. Replaced a fragmented set of third-party tools with a unified internal solution.
Integrated OpenTelemetry across 12 services. Built a custom log aggregation pipeline processing 500GB/day. Reduced mean time to detection for production incidents from 18 minutes to under 3.
End-to-end testing infrastructure built on top of Playwright, with custom reporters, parallel execution, and CI integration for a large React application.
Grew test coverage from 12% to 78% over six months. Parallelized test execution across 8 workers, cutting CI run time from 45 minutes to 9. Built a flakiness detection system that automatically quarantines unstable tests.