Machine Learning Data Platform: Working on Ads ML data infrastructure at scale for Large Recommendation Models (LRMs), focusing on long-sequence feature engineering, and training data pipeline optimization to support reliable and efficient model training.

Ads Training Data: Developed a modular Java-based batch training data framework for large-scale ads ML pipelines, reducing end-to-end runtime by 60% via Spark optimization and parallel processing.

Computational Advertisement: Built real-time data infrastructure for ad recommendation systems, improving CTR by 2% and enabling KPI-driven decisions with scalable backend services.