Overseeing the development, deployment, and monitoring of our acoustic classifiers across multiple ocean sensors.

- Developed and optimized ETL pipelines for processing vulnerability data across Salesforce using SQL- Reduced data processing times by over 50% through optimizing the MySQL cache size and redesigning tables to support greater scalability- Enhanced infrastructure and CI/CD processes by containerizing the platform with Docker and deploying via Terraform, Spinnaker, and Kubernetes

- Developed a real-time, collaborative retrospective tool to facilitate more effective and efficient sprint analyses - Utilized Lightning Web Components and JavaScript to design the UI and Apex to build the server - Optimized scheduled, high volume SQL transactions through batch processing

Assisted students with homework and projects pertaining to data manipulation and visualization (Numpy, Pandas), machine/deep learning (Scikit-Learn, Tensorflow) and probability theory at labs and office hours in the course DS100 (Principles and Techniques of Data Science).

We do contract work for companies in the Bay Area. Some of the projects I've contributed to include:Aptiv: - Consulted with Aptiv, a leading autonomous vehicle company, to develop a pipeline for assessing the safety of a driver's surrounding environment- Implemented a GRU network with PyTorch to predict nearby vehicle trajectories from autonomous vehicle data, based on the methodologies outlined in this paper - https://arxiv.org/pdf/1704.07049v2 Polymorph:- Collaborated with Polymorph, an ad-tech startup, to create predictive models for consumer ad click-through rates- Researched and implemented field-aware factorization machines in Python, achieving a higher F1 score compared to logistic regression and boosting models- Parallelized the training process with Dask, which reduced training time by a factor of 10

- Built an internal peer-to-peer employee rewards system with a private Ethereum blockchain using Solidity smart contracts- Implemented a Node.js REST API with web3.js that interacted with the Ethereum nodes on Google Cloud Platform

- Developed a crowd-sourced pedestrian/bicycle collision application for California’s Strategic Highway Safety Plan using HTML5/CSS3/Javascript and Leaflet mapping software.- Enhanced existing software tools by updating and debugging PHP and Javascript code.