• Lead the Realtime Application Team

• Collaborated with NYU’s Fenton Lab to design and implement a robust infrastructure leveraging the NYU High-Speed Research Network (HSRN) with the CoreLink protocol, successfully achieving a target data transfer speed of 100 GBps.

• Designed and developed a LSTM Machine Learning pipeline utilizing C++ and Tensorflow(Python), along with a real-time analysis system based on CalmAn and FLOLA, to process neuroactivity data from mouse brains, delivering live prediction results to the Fenton Lab via HSRN.

• Deployed the pipeline and associated models across a distributed Kubernetes cluster spanning seven NYU buildings, utilizing 14 NVIDIA GPUs to significantly optimize computational efficiency and performance.

• Worked on the Amazon Aurora Relational Database team, developing an AI agent and a Model Context Protocol (MCP) server for automatic ticket resolution.

• Developed the MCP server in TypeScript to interact with multiple internal operational APIs, databases, and other servers. Implemented various authorization methods, including the publicly known SigV4, and built comprehensive reliability engineering and region information resolution systems.

• Implemented the AI agent using StrandsAgent, enabling interaction with multiple AWS infrastructure

components, including Bedrock, internal git repositories, various logs, databases, operation tools, apis

• Built a knowledge base on top of OpenSearch Service, S3, and Bedrock.

• Managed the entire system with a well-written, fully automated pipeline using AWS CDK and Pipelines