• Cleaned up various L3 synchronization issues between switches, greatly improving internal testing results

• Played a key role in stabilizing switch firmware development for a new hardware line, which involved:

1. Collaborating with Core and Platform teams to clean up the compilation of different libraries

2. Providing continuous testing reports of the firmware image stability over the course of multiple months

3. Overseeing the merge of the team's separate development workflow into the main codebase

• Designed and implemented a robust system in C++ in Linux to accept, verify, and save JSON menu formats for use in voice ordering for restaurants. This was used to onboard 250+ different restaurants/businesses within the first year.

• Developed efficient Python scripts to convert and upload menu information to the C++ menu ingestion service.

• Created an experimental NLP fallback system in C++ to catch and process thousands of queries that aren't recognized by the initial pass through the NLU system.

• Helped renew SoundHound's voice AI contract with MasterCard and WhiteCastle for drive-through

food ordering. Helped transition the drive-through ordering project to be live-testable in a real

Chicago WhiteCastle location.

• Onboarded and managed four interns who all quickly became valuable contributors to one of the

company's core projects.

• Helped establish the restaurant team's agile development process by experimenting with mob programming, developing team style guidelines, and implementing a parallel git workflow for code pushes to production.

• Became the primary owner and developer for the C++ codebase responsible for handling food recipe queries.

Held office hours helping students to understand core concepts of operating systems, including threads, concurrency, race conditions, memory management, debugging, and more. Taught students how to complete projects that include: coding a thread library, coding a pager, and coding a filesystem.

• Designed and built the C++ codebase designated for handling food recipe queries.

• Utilized rule-based Natural Language Processing to efficiently and effectively categorize and address different user intents.

• Published the codebase to the production-level environment, allowing the Houndify platform (voice assistant) to correctly handle real user queries related to food recipes.