Used React Hooks to create stateful panels in the VS Code DOM for each extension’s scan results which updated based on event listeners living in useEffect hooks, avoiding memory leaks and matching the dynamic nature of the file searches to quickly and efficiently display information to the user

Highly optimized file scanning of a user’s installed extensions by using Node JS readstreams that would serve data in small asynchronous chunks as opposed to the default behavior which blocks code execution as a large single payload is prepared, improving speed efficiency by nearly 30-50% and reducing memory usage by 90% with larger files (1GB+)

Made use of the Unstructured API as a data retrieval source to parse PDFs and feed data to the LLM due to its accuracy and use of OCR (Optical Character Recognition) with cheaper costs compared to a product such as Amazon Rekognition

Developed an AWS serverless backend that interacted with AWS API gateway to make the code available online with API keys for security and access to rate limits for extra security and potential protection against spam/endpoint abuse

Led Test-Driven Development (TDD) practices in the team by devising comprehensive Jest and Vitest testing strategies for the entire stack, achieving over 92% test coverage and ensuring the successful functionality of frontend React components and Express API routes

Leveraged MongoDB to store and manage large volumes of order and product data, utilizing its flexible schema design and built-in sharding support to enable efficient scaling and high availability

Orchestrated Bitbucket pipelines for CI/CD, SourceTree for version control with a rebase focused git workflow allowing smooth PR merges and better enabling thorough commit audits by the development team

Led the creation of a chatbot written in JS on GCF (Google Cloud Functions) making use of Google Cloud Task Queues to stop duplicate events from being queued while preserving message order, streamlining the entire main trading process for the company and allowing CSRs to finish trades 70% faster

Co-led a year long TypeScript refactor of the codebase from JS due to the extreme value of type safety that is especially important in financial services, allowing us to eliminate and catch many bugs before even leaving local testing

Formulated methods of Service Decomposition, optimizing the Lambda functions by breaking up monolithic lambdas into smaller logical parts, cutting the deployment size that was originally near the 250MB maximum by nearly 30%

Spearheaded and led the creation and secure development of Bitcoin and Lightning wallets that incorporated unique methods of RBF batching to minimize network transaction fees

Introduced functional tests suites in Jest using database mocks and Selenium for E2E tests that would run through multiple potential order flows with ~70% coverage increasing code reliability and bug ousting within the lambda APIs and frontend