Rapid Prototyping @ The Forge- Designed, with close collaboration from stakeholders for iterative feedback/development loops, LLM- and NLP-based solution for alternative suppliers sourcing- Deployed scalable cloud-native, RESTful API-based prototype in condensed timeframe to meet critical business need- Architected automated database updates for up to 99% information relevance and reliability over time- Documented and presented system architecture diagrams, establishing robust framework for prototype to product pipeline- Trained linear regression model for employee-skill insight generation, increasing HR productivity

- Engineered a sequence-to-sequence alignment system between English and IPA with weighted finite state transducers (WFSTs) in Rust- Trained a model using the expectation-maximization algorithm with WFSTs and semirings to learn a set of English grapheme to IPA (G2P) phoneme mappings in Rust - Extended the open-source Epitran2 project with English grapheme-to-phoneme (G2P) training and alignment extraction

TA for CS 303E Elements of Computers and Programming | Technical Mentor for 100+ Students- Mentored students in Python fundamentals, algorithms, and software best practices- Built students' debugging proficiency through hands-on guidance- Strengthened students' coding independence through targeted feedback on code quality and efficiency- 95% of students successfully completed complex programming projects independently by course end

Software Engineer Intern on the Privacy Insights Team- Deployed automated system for machine learning (ML) embedding violation remediation using Python and SQL, reduced human effort by 95% and saved up to 20 hours of weekly manual effort- Engineered query API for budding internal privacy insights tool with Python and Hack, integrated API into automated remediation pipeline- Extended ML embedding violation data model in Python and Hack- Launched scheduled script in Bash to automatically process 1000s of database entries for API queries, introduced database logging for 100% coverage of creation/update/error events with SQL and Hack- Upheld quality and integrity of new and existing code by writing extensive unit and integration tests in Python and Hack, continued 100% automated test coverage