I architect fault-tolerant backend infrastructure built for scale and reliability.
- 🚀 Currently: Building distributed async job processing systems with Redis-backed priority queues, atomic Lua scripts, and recovery schedulers
- 🎯 Targeting: SDE-1 roles at product companies and mid-to-unicorn startups
- 🧩 DSA: 300+ problems solved across Graphs, DP, Backtracking, Trees, and Binary Search
Fault-tolerant job processing system sustaining 33+ jobs/second with tier-based priority scheduling, heartbeat monitoring, and automated recovery.
- Stack: Java · Spring Boot · Redis · MySQL · Docker
- Key Features:
- 🏎️ Priority Queue — Redis ZSET with atomic Lua scripts — PAID jobs always execute before UNPAID and PUBLIC
- 📉 Fault Tolerance — Reduced recovery complexity from O(n) to O(log n) via Redis ZSET range queries — two independent schedulers detect heartbeat expiry and max processing time breaches
- 🔒 Admission Control — Distributed capacity enforcement via Redis atomic counter — multi-instance safe without JVM-local synchronization
- 💀 Resilience — Permanently failed jobs routed to Dead Letter Queue persisted in MySQL after exhausting tier-specific max retries
- 💓 Heartbeat Monitor — Detected crashed workers within 500ms by atomically updating processing store scores on every pulse
Kafka-driven backend reducing link creation from 18s to under 20ms via async VirusTotal scanning across 90+ security engines.
- Stack: Java · Spring Boot · Kafka · MongoDB · Redis · JWT
- Key Features:
- ⚡ Threat Engine — Offloaded malware scanning to async Kafka consumer pipeline — reduced link creation from 18s to under 20ms across 90+ security engines
- 🛡️ Redirection Engine — Enforced 5 safety verdict states — from instant redirect to hard block — via hash-based routing
- 🔐 Auth — Structurally eliminated bot registrations by enforcing OTP-verified email at every sensitive operation
- 🔁 Abuse Loop — 3 OTP-verified community reports automatically trigger re-scanning, requiring zero admin intervention
Languages
Frameworks & Build
Databases
Caching & Messaging
Tools
I'm always open to collaborating on open-source projects or discussing backend architecture.
