Build - I designed AWS Solutions by talking to customers, drafting architecture diagrams and building proof of concept code which were published to the AWS Solutions Library (https://aws.amazon.com/solutions/).Secure - I reviewed over 5 AWS Solutions Guidance created by peers to ensure that they meet AWS Content Security Standards.Innovate - Reduced the burden of publishing Solutions to the AWS Solutions Library by automating time consuming tasks. Built customized GenAI solutions that provided immediate content quality feedback to builders.

Provided vision and guidance for the new trademark search tool (https://tmsearch.uspto.gov/search/search-information) - a cloud based scalable solution that leverages server-less offerings.

Created AI based proof-of-concepts that address the key mission of USPTO - Searching for relevant patents and trademarks.

My Ph.D. thesis work builds on the Polyscheme Cognitive architecture which is able to combine various modes (language, vision, etc.) through a focus of attention mechanism. I built the probabilistic inference framework that accommodates indefinite objects which is essential for parsing PCFG. Jointly reasoning over syntactic and semantic constraints in natural language processing applications is shown possible through this approach.

Led research on a project that reduces the error rate of Automated Speech Recognition systems by using current events to provide context. Reported directly to the VP of Research.

Studied in Prof. Patrick Winston’s Genesis lab. Worked on the Natural Language Understanding framework - Genesis bridge and implemented word sense disambiguation using self-organizing maps.