At Syana, I built an AI-powered customer-retention platform that’s live at 15 automobile dealerships. Its call-intelligence pipeline transcribes service calls and uses LLMs to score quality, while a priority engine ranks customers for daily follow-up; paid-service retention rose by up to 11 points over 18K+ calls.
At Dream11, I architected video ingestion and player-identification pipelines for SportsAI, bringing processing in-house and reducing video turnaround time by 83%. The player pipeline processes more than 100 match videos daily and cut manual cropping to a quick review step.
I also worked on Dream11’s leaderboard platform, scaling it to 170M RPM and 170M entries while reducing IPL winner declaration time from about seven minutes to about two. Earlier, I migrated fantasy-point calculation for four sports into a Java Vert.x microservice that sustained 45M RPM at peak.
