Context
Rajasthan Royals, an IPL franchise with a 12M+ fan base, needed player performance evaluation and scouting workflows over a warehouse of millions of records. I owned the project end to end — infrastructure, API design, authentication and deployment — and communicated directly with their data analysis team.
Query performance
I built a REST API layer over a BigQuery analytical warehouse using multi-CTE drilldown queries and in-process aggregation, cutting query response time from roughly 1 second to 0.65 seconds.
The best-eleven algorithm
A two-pass greedy composition algorithm enforcing the IPL's overseas-player cap and salary-cap tracking, resolving season and competition context through a year-walk query with no hardcoded year.
Deployment and security
A full CI/CD pipeline — Docker to AWS ECR to AWS ECS via GitHub Actions — reduced deployment errors by roughly 30% after automation. Access used dual-token authentication (a JWT access token plus an HttpOnly refresh cookie) behind an RBAC middleware pipeline, backed by two independently isolated GCP service accounts for least-privilege IAM.