Ellipsis reviews code and fixes bugs. It's used by 100+ companies and installed in 19K codebases. Free 7 day trial available.

Lead several efforts to improve MLOps in service responsible for forecasting spend of AWS accounts. Efforts related to experimentation, A/B testing, measuring lift, collecting ground truth data, and applying a proprietary forecasting model in low latency environments.

Making technical interviews more like day-to-day work for software engineers.

Applying ML to low latency, distributed systems in the Advertising space. Led efforts to pipe campaign/advertisement/creative data into a demand side platform that served impressions at <100ms. Efforts led to 4% increase in CTR for first party display ads.

My first role (July 2016-July 2017) was as a core developer of a AWS data pipeline. I handled the creation and enhancement of a multi resource data movement pattern, paying special attention to the strict security, scalability, and best practice adherence mandates of a financial tech company. I used Python, EC2, S3, Lambda, Redshift, SNS, Cloudwatch, along with several homegrown solutions relating to Metadata Management, Data Quality, ISRM, and DevOps. During my time on this team I wrote a white paper on automated testing using GIVEN WHEN THEN statements and gave a few talks on it. Additionally, my experience engineering on the AWS cloud netted me several opportunities to consult with other Capital One development teams.My second role was as a founding developer for a novel Capital One product. Developed containerized microservices on AWS using NodeJS. Used RDS, PostGREs, S3, Kubernetes, ECS, and Lambda.

Published (https://arxiv.org/abs/1511.00651) on the significance of Stellar Mass and the Dust Attenuation Curve. Work involved creating a software package capable of: modeling stellar populations via various parameters, interpreting pure observational measurements, comparing various models to interpreted observations, transversing parameter space in effort to match model to observations, converging on estimate of galactic model parameters, interpreting posterior distribution of random walk process.Software successfully applied to data from 42 astronomical observations and run on 239 individual galaxies. Data science efforts culminated in 2 publications and 3 undergraduate research poster exhibits.