I design and ship production backend systems — distributed architectures, event-driven pipelines, multi-tenant platforms, and data-intensive services. Not tutorial projects. Real systems under real load.

TrueDialog — 28-service event-driven messaging platform on Azure (Service Bus · Event Hub · SignalR) processing 100K+ contacts
SuprSwitch — Distributed ETL engine with pluggable adapters for zero-data-loss migrations across 8 CRM platforms
Qezir — Live SaaS scanning 90+ sources in parallel with SSE streaming and distributed rate limiting
FastAPI Refactor — 15,000 → 3,500 lines, 4 days, zero production regressions

Dec 2025 – Apr 2026

Built the backend for a multi-tenant CRM asset deployment platform. Designed dependency-aware orchestration with idempotent distributed execution, retry logic, and progress tracking across Cloud Tasks, Firestore, and MongoDB.

Multi-tenant Cloud Tasks Firestore MongoDB OAuth Idempotency Distributed Systems

Dec 2024 – Dec 2025

TrueDialog — Designed architecture for a 28-service event-driven enterprise messaging platform:

• Azure Service Bus for reliable queuing, Event Hub for high-throughput streaming
• SignalR for real-time delivery tracking across 100K+ contacts
• Dead letter queues, retry mechanisms, bulk scheduling, webhook processing

SuprSwitch — Architected a distributed ETL engine for enterprise CRM migrations:

• Pluggable adapter pattern across HubSpot, Salesforce, Zoho, Dynamics, Zendesk, Pipedrive, Freshsales
• Cursor-based incremental sync, schema mapping, association rebuilding, idempotent upserts
• 67K+ record migrations with zero data loss

FastAPI Refactor — Rewrote a 15,000-line undocumented production codebase:

• Reduced to 3,500 lines with typed models, central validation, structured error handling
• Production deployed in 4 days, zero regressions

Azure Service Bus Event Hub SignalR Redis CosmosDB Terraform ETL OAuth

Live SaaS · 50+ daily users

Scans 90+ external sources in parallel to return brand availability, security posture, and social handle intelligence. Built with distributed concurrency, Server-Sent Events for real-time streaming, rate limiting, and multi-factor scoring.

Node.js SSE Distributed Workers Rate Limiting Multi-source Extraction

1st Prize · Selected from 44,000+ teams nationwide

AI-powered government job recommendation platform with AR integration, built for the Ministry of Skill Development.

TypeScript React Native Next.js Node.js MongoDB AR