● Created a Python / React prototype for a low-code ETL tool that generated a $3 million investment and led to a company initiative to make it a core product offering
● Led and mentored a team of ten engineers to transform ETL prototype into a fully functional product
● Established patterns for our team for the development, testing, and deployment of code on multiple projects. This includes patterns for architecture, unit & integration testing (local & CI/CD), building/deployment of images, microservice communication, authentication, multi-tenancy, clean code, development tooling, and git branching
● Wrote multiple microservices spanning ETL, ML inference, task scheduling, & authentication/authorization
● Created algorithms to extract matching rules from reconciled data to drastically accelerate client onboarding from competitors
● Wrote >90% of machine learning codebase allowing clients to perform preprocessing, training, and inference on tabular data. Supported models include both supervised (boosted trees, neural networks) & unsupervised models for anomaly detection (autoencoders, isolation forests)
● Utilized Shapley Additive Explanations (SHAP) to enhance the interpretability of ML models, employing global explainers to uncover insights into models' overall decision-making process and local explainers to provide clients with per-record explanations for model predictions
● Created Slack CLI to manage AWS resources directly in Slack (Used over 24000 times saving 2000+ developer hours) & converted this tool to a LLM chatbot to help our team manage resources using natural language
