• Embeddings, Search, Ranking, Retrieval, Graphs, and Evaluations of ML Systems

• Architected a rewrite of the in-house graph database on PostgreSQL; implemented sharded UUID scheme to encode the data model, ensuring records reside on the same memory pages and reducing query timeouts by 25%

• Performance tuning, Distributed Systems, Elastic Search, ingestion pipelines

• Python, Sagemaker, Celery, Django

• ML Systems Engineer working on workflow orchestration

• ML Platform, ML pipelines, Kubeflow

• Shipped several features that improved the scalability, reliability, and observability of the platform.

• Build, CI/CD, Distributed Systems

• Worked on Novelty Detection and adaptation in open-world Multi-agent Environments funded by DARPA.

• Developed an agent which could detect novelties in the context of the Monopoly game without having any pre-notion of the novelties.

• Achieved 70.3% accuracy in phase2 Evaluations by USC which was the best among all other universities participated.

• TextFooler on GPT-3, Gemini implementation on HumanEval

• Applied NLP in language models like GPT3 and BERT using Logical QA dataset and SNLI

• Working in the Compliance & Risk Platform to report fraudulent transactions to regulators.

• Worked in cross geo deployment and migrations of customers across geographies.

• Enhancement of platform service features, detecting leaks, and tuning APIs & queries to reduce response times.

• Migrated platform codebase and data to China with 0 downtimes and received spot award for completing it.