At adidas, I lead the development of a Real-Time Demand Intelligence Platform that brings Multi-Agent AI into demand forecasting, buying, and planning. I also engineered a natural-language interface for buyers and planners to explore production data, access insights, and take actions.
I led the architecture and deployment of an enterprise multi-agent AI platform that eliminated a one-week manual Excel workflow for demand forecasting. I also coordinated four concurrent initiatives across Data Engineering, MLOps, and Agentic AI with a cross-functional team of 8–10 engineers.
At adidas, I orchestrated the migration of more than 100 PySpark jobs from AWS EMR to Databricks Lakehouse, achieving a 500% performance improvement. Earlier, I guided demand forecasting and supply optimisation solutions, including the migration of legacy Alteryx and Exasol logic to AWS EMR PySpark.
At S&P Global Market Intelligence, I modernised financial data processing by migrating more than 50 jobs from SQL Server to Apache Spark, achieving over 400% performance improvement. I also developed Spark-based Elasticsearch indexing and moved workflow orchestration from Oozie to Apache Airflow.
