Apr 2025 – Present · 11 mos

• Built the industry’s first real-time game streaming platform with sub-second latency.
• Modernized 100+ CI/CD pipelines across 3 teams, enforcing governance and improving release reliability.
• Delivered reconciliation systems ensuring 100% accuracy across millions of tournament data points.
• Cut inter-game delays by 40% and boosted games per user by 5% through gameplay optimizations.
• Led and mentored the 7-member Core Gameplay team responsible for all Rummy game brands, driving delivery of flagship features at scale.
• Introduced AI-driven PR reviews and workflow automation, improving code quality and developer velocity.

Aug 2021 – Jun 2025 · 3 yrs 10 mos

• Scaled tournament system from 50K → 250K participants using Redis (Lua, Pub/Sub), BeeQueue, and Socket.IO, enabling continuous gameplay and powering the All India Rummy Grandmasters Challenge.
• Built a real-time feedback system for 500K concurrent players with Redis, MongoDB, and AWS SageMaker, improving decision accuracy by 5%.
• Implemented real-time anomaly detection with Flink, Kafka, and Elasticsearch, cutting incident response time by 80%.
• Developed auto-scaling for Flink clusters, now the backbone of all data pipelines, improving resource efficiency and reliability.
• Optimized matchmaking and large-state processing in Flink (RocksDB, retries, broadcast state), reducing latency and improving fault tolerance.
• Reduced Redis downtime from 15 min → <1 min via ioredis customization, retry queues, and failover strategies, and improved SLA compliance with MongoDB TTL fixes and automated recovery workflows.