Who Am I?

Technical Experience

Downloadable Version

APrime Technology Senior Software Engineer 2022-Present

During my time at APrime, I was a consultant that focused primarily on data engineering and devops. During my time consulting, I worked with companies ranging from pre-seed funding all the way through Series B companies across industries including ETF investmenting, usage-based billing platforms, benefits reimbursal, and many others. For larger companies I worked embedded in existing engineering teams and for smaller companies I worked more greenfield suggesting and implementing best practices. Because of this, I worked with a wide-range of different technologies and languages that addressed problems such as running a a GCP to AWS cloud migration, full private VPC migration, in-place kubernetes upgrades, custom Go based migration tool, full end-to-end machine learning pipeline using Prefect and R, migration of AWS Step Functions to Dagster, deployment of Argo Rollouts and ArgoCD, and others.

New York Times Senior Software Engineer; Machine Learning Infrastructure 2021-2022

My title at the New York Times is a bit wordy, but I, essentially, worked on the machine learning platform team that was in charge of serving all recommendations to end users. This includes things like recommending articles that a user might like, generating a top-5 of articles to read for a user, and creating a newsletter to drive users to the frontpage. I focused mostly on revamping the CI/CD process during this time to reduce deploy times as well as migrating our kubernetes cluster to a service mesh to allow for shadow and canary deployments.
Our endpoints received thousands of requests per second and we had a SLA of around 100 milliseconds to return a response to the internal gateway API. Given this, all of our models were written and experimented in Python but then subsequently translated into Golang to provide the best latency possible.

Bark Data Scientist to Machine Learning Engineer 2019-2021

I worked in a number of different roles at Bark, but I started as a data scientist and worked my up to a machine learning engineer. This meant that I started off working on our recommendation engine that we used to try and get users to add extra products to their monthly shipments. But, this also spun off into projects including trying to predict an estimated time of delivery and running survival analysis on our subscribers. Over time, though, I need to work further and further back in the stack and started shifting my focus from just the model but to the whole model lifecycle (i.e. how do we deploy, monitor, and update models safely). This involved standardizing all of our model training using KubeFlow (on GCP) to deploying models with TensorFlow Serving.
Of course, there were a number of other projects that came up, as is common in any startup. I migrated our MongoDB instance to Amazon Redshift, did some causal analysis to try and determine the effectives of our interventions in preventing subscriber churn, improved our sampling weighting to better capture a stratified random sample for our NPS survey, and creating a microservice to deploy our predictions rather than running everything as a batch process.

Trendalytics Software Engineer; Machine Learning 2018-2019

We were focused here on doing lots of image segmentation, image processing, and time-series analysis to try and predict upcoming fashion trends for clients. Most of my time here was spent on improving our implementation of the Yolo algorithm we were using for image segmentation and also looking at how we can combine a bunch of different model outputs (i.e. keyword search trends, NLP text extractions, and image vectors) into one singular usable score for our clients to action upon.

Institute for Health Metrics and Evaluation Data Analyst, Machine Learning Data Specialist 2016-2018