• Developing an A/B/n testing framework integrated with LLM (Code Llama), vectorDB, and graph DB to map RDBMS relationships, enabling intelligent Snowflake queries for enhanced data-driven decision-making.

• Deployed a scalable real-time data pipeline for drive classification, leveraging advanced spatial-temporal feature engineering to enhance ML model accuracy by 10%.

• Led a 30-member engineering team in developing a real-time profiling system for BFSI, processing 600K events/sec and 25TB/day, ensuring 99.99% availability.

• Implemented a Union-Find-based graph deduplication system, clustering similar data entities in O(α(n)) time complexity, achieving 50ms lookup speeds and reducing storage/operational costs by 50%.

• Designed and deployed facial recognition and fuzzy matching pipelines, increasing identity verification accuracy by 15%.

• Pivoted from AWS to Kubernetes in 3 months, cutting setup time and standardizing deployments across financial institutions to meet evolving business needs.

• Collaborated with PMs to prioritize and deliver features that drove a 20% increase in MAU growth.

• Built and optimized ETL pipelines handling 3GB/day of data, improving processing efficiency by 30% using AWS Glue, Kinesis, and Step Functions.

• Implemented incremental data ingestion pipelines, reducing compute costs by 30% via delta calculation techniques.

• Developed Kimball-modeled data warehouses in Snowflake, improving analytical query performance by 5%.

• Automated cloud infrastructure provisioning with CloudFormation, accelerating deployment cycles by 1.5x.

Full-stack development and Linux shell scripting

• Worked on full-stack development for a marketplace application using Angular 8 and NodeJS.

• Worked on full-stack development for an accelerator product using Angular 8 and Python.

• Worked on Linux shell scripting to check best-practice implementations in AWS.