During my time at MORSE, I've been a part of numerous projects on both the Data Intelligence and Test Range Automation portfolios.
My first project was a wrapper library around Matplotlib to simplify and standardize AI T&E report generation. I started by refactoring the numerous, diverging plotting functions to a few extensible and reusable functions as well as creating new plotting functions for visualizing T&E metrics. This library is in use both internally for all T&E projects at MORSE and also in the Joint AI Test Infrastructure Capability (JATIC) program under the Chief Digital and Artificial Intelligence Office (CDAO) in the DoD.
Following my work on JATIC, I was part of a novel, AI-enabled toolkit for simulating vehicle and sensor interaction in real-world environments. The toolkit was developed using Python for configuration and analytics, a C++ compiled library for simulation runtime, and Typescript/React for a simple web UI. After 6 months of developing this toolkit from the ground up, I was promoted to Task Lead, leading a small team of engineers by handling task management, running standups and retros, and working with Chief Engineers for sprint planning and requirement gathering (all while actively developing new features for the toolkit). In between this project's development, I worked on another program developing and maintaining T&E pipelines for vendor-provided models. I was in charge of moving our pipeline orchestration from an internal tooling library to Airflow, enabling simpler deployment capabilities and reducing the number of developers needed to maintain pipelines.
Currently, I am working on the Test Range Automation portfolio as a full-stack developer on a few products supporting efficient data management at ARMY test centers.
Software Used: Python (Pydantic, matplotlib, NumPy, Pandas, Pytest, OpenAPI), C++ (CMake, pybind11), Typescript, React, Airflow, Gitlab CI, Jupyter, Docker
I interned at Salesforce on the Database Observability Team. My intern project for the summer was to migrate the health reports database from Apache Hive to Apache Iceberg. Following the migration, the average latency for querying the health reports database was reduced by 65%!
Software Used: Java (JUnit), Trino, Kafka, Apache Hive, Apache Iceberg, S3, PostgreSQL
I was a part of the MLOps team at PathAI, managing the internal stack used for the creation, execution, and monitoring of machine learning workflows. One of my first tasks was the design of a testing methodology that would identify if local Kubernetes pods were able to perform file locks and write to a shared mount drive. My biggest contribution to PathAI was the implementation of a "high-priority" flag that could be enabled for machine learning workflows ran on our Kubernetes cluster. Using this flag, workflows were given heightened execution priveleges, allowing them to run to completion even if the cluster was being flooded with requests. As this was a very powerful option to give workflows, I also created a Slack bot that would notifiy ML and MLOps managers for when high priority workflows were launched and who was resposible for their execution. Additionally, I created a DataDog dashboard that would provide allow resouce usage monitoring as well as tracking how many high priority workflows were being ran and for what tasks.
Software Used: Python (Boto3, NumPy, Pytest, PyTorch, Flask), Gitlab Ci, S3, Kubernetes, Jupyter, Docker, Vue.js
I started at MORSE working on a small data engineering team. After a few months, I started the co-development of a library made for efficiently building machine learning pipelines. Once we finished the library, I transitioned to leading a team of 4 engineers in the design of a distributed AI networking project. With the help of my team, I wrote system requirements, identified the project's scope, and outlined a test plan for when the product would be implemented. The finished design was meant to be presented to key stakeholders that would decide whether or not MORSE would be able to implement the system fully. As part of this presentation, I created a number of animation using Manim to help better illustrate the system's capabilities and functionality.
Software Used: Python (PySpark, Pandas, NumPy, scikit-learn, Pytest, Matplotlib, Manim), Spark, Jenkins, Docker, JIRA
At UEI, I worked with the Sales and Applications teams creating internal software soltuions. Towards the end of my time at UEI, I took over the initial testing of UEI's first Python library.
Software Used: Python (tkinter, OpenCV), Java (JSwing), VBA
Algorithms & Data, Object-Oriented Design, Networks & Distributed Systems, Computer Systems
Natural Language Processing, Machine Learning & Data Mining
Digital Logic Design, Embedded Design, Circuits & Signals, Electronics
Calculus 3, Differential Equations and Linear Algebra, Probability and Statistics, Graph Theory