Anshika Marmat

Developed and scaled a high-throughput platform for large-scale ads automation across online marketplaces. Designed a three-layer ads automation flow engine using Amazon APIs that re-balanced budgets and restructured 200+ campaigns. Deployed backend services on EKS using reserved EC2 instances, decoupled components with SQS, implemented Redis for caching, and tuned DuckDB–Parquet queries. Integrated an ML pipeline using AutoML that trains on data from DuckDB every 6 hours. Built an event-driven Python pipeline using AWS Lambda, Step Functions, and S3 to process 1.2M records daily. Developed a Slack alerting engine using Python Bolt & AWS Lambda. Also developed an end-to-end Prompt Engineering platform using Python, FastAPI, and PostgreSQL, implementing Factory Design Pattern and Dependency Injection.

Tested 50+ workflows across floorplans, invoicing, and payment modules in Salesforce-based ShowCycle, ensuring 95% test coverage. Reported and documented 25+ bugs and collaborated with a six-member team. Documented a 30+ page user manual, reducing client onboarding time by 40%.

Increased consumer engagement and retention through effective social media strategies for 8 leading brands across agriculture, tyre, and FMCG sectors, boosting engagement by 25% overall. Created product launch campaigns including BTL and digital for 5+ clients in the agriculture sector.