Building ML Platforms at Oracle. Built distributed training framework for Tensorflow and pytorch Pytorch workloads. Also building model evaluation framework for LLMs and speech models. Also built assisted labelling capabilities of using ML models for NER and face detection for auto-labelling in OCI Labelling solution. Built a platform for evaluating ML models on GPUs and a metrics store with reporting capabilities. Solved Data challenges at MakeMyTrip for deploying Machine learning models to production. Computing features for model training and real-time serving using stream processing systems like Kafka-streams, Spark streaming and batch processing pipelines in spark with airflow. Acting as a consultant for all things engineering for the Data science team be it data quality or low latency high scale system design Worked with Inmobi for over 4.5 years. Worked on a complex Ad-serving system that serves a relevant Ad for each request that the system gets. Owned the Spark pipelines that provide intelligence to this system for dynamic pricing. At Microsoft, performance tuned our multi-tenant Dynamics CRM backend APIs and developed mobile web services hosted on Azure (Microsoft Cloud-based solution)