• Developed full stack application leveraging LLMs and Speech Models to conduct voice surveys in lieu of traditional call centers (Next.js/Python/LangChain).• Lead prompt engineering efforts including schema design, A/B testing, tracking and evaluation

• Spearhead connected data product integrations.• Architected and implemented a scalable, real-time entity-resolution application (Go Kafka Consumer with an ElasticSearch backend). • Developed, deployed, and tested a dynamic pricing application (FastAPI) utilizing contextual bandits (VowPal Wabbit) resulting in a $60-100k monthly increase in revenue.• Engineered and scaled PyTorch based missing data imputation.• Improved CI/CD workflows (TeamCity/Ansible/Kubernetes/Kustomize) and cloud integrations (DataDog/AWS).

• Led a data science/engineering development team focused on building custom data-driven solutions in the advertising technology space.• Architected a distributed machine learning workflow management application built using Spark, Kubernetes, Scikit-learn, H2O, Django, PostgreSQL, and Vue.js.• Played a key development role in several additional data products using technologies ranging from data visualization (React.js, D3) to Bayesian modeling (STAN, JAGS) to web scraping (Scrapy).• Led SRE efforts including migration to cloud-based technologies, containerized deployments (Docker, Kubernetes, Terraform, GCP, & AWS), and implementing CI/CD.

• Managed a team of quantitative analysts responsible for designing, implementing, and analyzing surveys, randomized experiments, and predictive models.• Developed a predictive modeling pipeline using Luigi and the Python scientific stack (numpy, pandas, scikit-learn) capable of predicting turnout likelihoods for all of the nation’s 180+ million registered voters.• Conducted cost-benefit analysis of using different advertising mediums (mail, digital, television, door-to-door) for several nonprofits and political campaigns.• Designed phone/digital surveys and randomized control experiments.