Work
Every project tells a story. Here are the systems I've built, the problems I've solved, and the lessons that shaped how I engineer.
Client Work
Production systems I've architected and delivered for companies.
Patient Management Platform
A comprehensive patient-management platform serving modern clinics, hospitals, and telehealth providers. Handles appointment scheduling, real-time patient data analytics, and secure health records at scale.
The client needed a HIPAA-compliant platform capable of handling thousands of concurrent patient sessions with sub-second response times, deployed across multiple geographic regions.
Delivered a NestJS microservices backend on GCP Kubernetes with Firebase integration, achieving 99.9% uptime and supporting real-time patient data sync across 3 regions.
Industrial IoT Monitoring
A real-time industrial IoT asset monitoring platform that tracks machinery, pipelines, and hardware at scale using MQTT-based messaging and cloud infrastructure.
A manufacturing plant needed to ingest and process millions of sensor data points per day from hundreds of IoT devices, with anomaly detection triggering alerts within 5 seconds of occurrence.
Architected an MQTT-based pipeline using HiveMQ on GCP. The system processes 2M+ daily sensor events with predictive maintenance alerts, reducing unplanned downtime by an estimated 30%.
Fitness Tracking Platform
A personalized workout tracking application with goal setting, activity logging, and progress analytics. Designed as a virtual personal trainer with real-time updates and intuitive visualizations.
The fitness startup needed to go from concept to launch in under 3 months with a small budget, requiring a fast iteration cycle and a tech stack that could scale with user growth.
Built and shipped the full platform — MongoDB + Express backend, React frontend — in 10 weeks. Supported 5,000+ users in the first quarter post-launch with zero downtime.
Distributed Queue System
A microservices-based queue management platform leveraging gRPC and Protocol Buffers for high-performance inter-service communication. Built for companies handling large-scale event-driven data flows.
The existing queue system used REST-based communication between services, creating latency bottlenecks at 10K+ concurrent messages. The client needed sub-10ms inter-service latency at scale.
Replaced REST with gRPC + Protocol Buffers, achieving 4ms average latency between services. The system now handles 50K+ concurrent messages with fault-tolerant reliability using Go-based microservices.
The One That Didn't Work
Not everything ships to production. Here's what I learned from the projects that didn't make it.
TeamRef
I spent months building a complete SaaS platform for employee referrals — multi-tenant, Slack bot, kanban pipeline, Docker deployment, and a full marketing site. It didn't take off commercially. But I shipped something real, and I learned more from this failure than from any success.
Could I build and launch a complete B2B SaaS product entirely solo? I wanted to test every skill — full-stack development, DevOps, product design, marketing, and sales — all at once.
I shipped a fully functional product with 99 commits across 3 months: multi-tenant SaaS, Slack bot integration, kanban referral pipeline, landing page, and Docker deployment. But I validated nothing before building. I spent 90% of my time on code and 10% on talking to customers. I built something technically solid that nobody asked for. The product works. The business didn't.