At Info Edge (AmbitionBox), I resolved approximately 200 daily cache-refresh API timeouts by addressing thread-pool exhaustion and introducing batched jobId-based bulk APIs, cutting timeouts by approximately 93%.
I also re-engineered employee-benefit extraction with rule-based NLP using OpenNLP, adding synonym and negation handling to improve extraction by approximately 40% across 12Kâ15K daily reviews.
At Paytm Money Limited, I developed backend systems for fund-transfer retries and transaction recovery using Redis and Amazon SQS. I also built a Kafka-driven pipeline for bulk mandate failures that resolved incorrect SIP linkage for 100K+ dormant mandates.
My projects include a full-stack fitness platform and a YOLOv8-based vehicle detection system. For the ANPR project, I evaluated EasyOCR, KerasOCR, and Pytesseract, achieving 97% vehicle-detection accuracy and 89% OCR accuracy.

