Gyanendra Parmaar

I’m a Software Engineer at Myntra, focused on building production ranking and real-time data systems that improve both user experience and system performance. I enjoy owning end-to-end pipelines—from architecture through reliability and experimentation.

At Myntra, I architected a hybrid ranking API pipeline that merges long-term preferences with NRT session data, delivering a +1.67% CTR lift for Homepage ads during A/B testing. I also reduced end-to-end P99 latency to less than 100ms for a candidate set of 200k+ ads.

To support ranking logic with fresh metadata, I engineered a high-throughput Kafka-to-Aerospike and Aerospike-to-Kafka bridge for real-time synchronization. I also built an Advector Service that generates vectors for 200k ads per week in under 1 second, with automatic reconciliation to maintain high availability during failures.

Beyond core backend work, I develop automation using AI for business workflows and creative generation—using a VLM-driven agentic framework for ad-creative workflows to reduce product occlusion to under 10%. I bring a strong ML/AI mindset alongside solid systems engineering, from microservices and REST/gRPC to scalable infrastructure.