Work on the Machine Learning Platform Team to develop software tools that make the Machine Learning Engineers' lives easier.

• Built Jupyter notebooks on Supervised Learning to develop Codesmith’s Machine Learning curriculum, leveraging Jupyter’s literate programming and interactive computing paradigms to create an interactive learning environment.

• Implemented Jest tests for Codesmith’s web server controllers, helping to ensure the maintainability and reliability of complex asynchronous requests and interactions with a SQL database for a quickly updating codebase.

• Mentored younger engineers on production level software projects, organizing code reviews to keep their teams up to date, guiding standups to keep the teams on track, and providing technical assistance on their codebases.

• Built a Node.js tool utilizing a Node/Express server and a React based GUI to organize GraphQL queries to an endpoint, gather response timing data for these queries over time, and interactively visualize the timing data.

• Used Yargs to augment the command line interface, creating a descriptive help menu, extensive option support, aliases, and easing future development by laying down an easily maintainable and extensible file structure.

• Utilized D3’s data visualization API to provide interactive and dynamic visualizations of GQL response timing data.

• Built a GraphQL server interfacing between endpoints with varying response times to test our timing gathering tool.

• Configured Webpack for easy development with hot module reloading augmented with proxy server, transpilation of JSX, TSX, and Typescript, and used webpack to uglify and minify build for compact production level code.

• Migrated React components to TypeScript (TSX) to circumvent potential typing errors as the codebase develops.

• Used Jest to implement a testing suite for the Command Line Tools to ensure that core functionality is preserved as the project grows in size, improving the extensibility of the project and streamlining the development of new features.

• Used React to build modular and reusable components with independent business logic, separating different user-facing functionalities via conditional rendering, and utilizing React’s virtual DOM to optimize performance.

• Set up an Express server with extensive middleware to handle data-flow from the front end to the backend in a clean way, reducing cognitive overhead and maximizing ability for future developers to iterate on the project.

• Built SQL database to store known user data and user input, leveraging SQL’s ability to query data based on any feature to easily retrieve structured data for future analysis, and relying on SQL’s ACID compliance to ensure that the data will be preserved even in the event of emergency shutdown, maintaining the accuracy of the data analysis.

• Used React Hooks to add new functionality to an existing codebase, and to manage increasingly complex state in a SPA with nested functional components, keeping front-end logic more modular and clean for future developers.

• Used React-Router to build a responsive and user-friendly SPA, allowing users to keep track of browser history and route between pages without having to add complexity to the backend, creating a smoother user experience.

• Integrated a third party API to generate geolocation data to provide to the frontend for interactive visualization of parking spots, optimizing user experience.