• Evaluated LLM outputs using Python and internal APIs for factuality, helpfulness, and latency.

• Flagged quality gaps, leading to +22% factuality, +4.2% helpfulness, and -5% latency.

• Ensured alignment with user intent and minimized hallucinations through in-depth response evaluations.

• Maintained data integrity via QA protocols and peer review of analyst assessments.

• Contributed to prompt design, regression analysis, and refinement of tools and workflows.

Health Hustlers is a comprehensive full stack application for tracking workouts and managing meal intake.

Role:

• Collaborated with a team to design the frontend and backend flow using wireframes to create an intuitive user experience.

• Implemented secure authorization for both the frontend and backend using JSON Web Tokens, ensuring appropriate access based on user roles.

• Deployed the application on Galvanize Cloud and utilized GitLab CI/CD pipelines for efficient code review, merge requests, and passing pipeline tests.

Wardrobify is a user-friendly app that serves as a virtual wardrobe for storing hats and shoes, enabling users to effortlessly explore and navigate their hat and shoe collections.

Role:

• Used React for the front-end and Django for the back-end, enabling users to efficiently sort and manage their collection.

• Utilized React forms to enable users to easily add new hats and shoes to their wardrobe, improving the overall user experience and streamlining the data input process.

• Implemented robust functionality for deleting shoes and hats, enhancing the app's versatility and allowing users to easily manage and update their wardrobe collection.

AutoZone is a web application for managing automobile dealership—specifically its inventory, service center, and sales

Role:

• Created poller to receive dynamic inventory data into a service microservice for vehicle information

• Containerized microservices intercommunication using docker

• Implemented web forms with React to manage automobiles sales, repairs service, and inventory

• Designed and tested backend APIs to manage data requests from React service

Conference GO is a conference management solution, streamlining registrations, scheduling, and managing the presentations and locations of the conferences.

Role:

• Developed and designed an conference management application utilizing Django, Python, CSS, and JSON

• Integrated third party APIs to include a photo based on the inputted location and incorporate weather data for the users

• Streamlined accessibility and improved maintainability of the application by refactoring from a monolithic to microservices based application