• Architected MLOps infrastructure for training & deploying 30 ADME-Tox models to all internal drug discovery projects.• Research engineering work focused on evaluating & benchmarking new artificial intelligence / machine learning (AI/ML) methodologies for small molecule generative modeling & high throughput proteomics screen analytics.• Championed modern software development & product practices on internal and cross-functional project teams.

• Hybrid Team Lead & Senior Contributor role reporting directly to the CEO.• Developed high-performance, big data analytics pipelines in Python on a unique industry-leading platform for time series analytics.• Presented platform demonstrations to high-value clients, and presented my projects to the national power engineering community.• Mentored team members and steered data science strategy in leadership meetings.

• Senior Data team member and mentor in machine learning, statistics, data analysis & visualization, programming, & git.• Developed Bayesian forecasting features for an operations model used to understand clinical trial recruitment performance.• Led a cross functional project to develop and productionize a machine learning-based patient lead ranker that led to a 9% increase in patient reachability.• Created an NLP-powered command line tool to extract information from physician notes to match patients to clinical trials.• Contributed to internal and cross-functional engineering projects regarding codebase refactoring & alerting.• Developed metrics to measure data quality on a unique healthcare data model.• Built time series models (SARIMA, recurrent neural networks, Bayesian structural time series, moving averages) for sensor data and marketing forecasts using Tensorflow and other libraries.• Technologies used: Python data stack, SQL, Looker, dbt, Redshift, & others• Wrote 2 blog posts for the TrialSpark Engineering blog.• Invited speaker to the American Physical Society’s annual meeting 2019, industry track.

• Managed the Data team’s work regarding (i) analytics projects to support Leadership, Sales, Account Management, Marketing, and other internal teams, as well as (ii) infrastructure projects that scaled query capabilities and storage to over 5.4TB of data in AWS S3/Athena/Redshift.• Invited speaker & panelist at the EGG 2017 Data Science Conference. Also presented at the Dataiku Meetup as well as various internal presentations.• Made various individual contributions

• Developed & productionized machine learning models (NLP) for article recommendation (increasing clickthrough rates by >3x) and job classification (allowing the job board to scale from 1,000s to nearly 100,000 live jobs) in scikit-learn.• Created lead scoring models, statistical reports & dashboards, and productionized backend integrations with applicant tracking systems.

• Wrote and maintained multiple ETL integrations with public data APIs in Python for StoryPilot, a data-driven web app created to help media makers better understand the social impact of over 500 documentary films in 16 social issue categories.• Contributed data analysis (social network analysis & others), codebase refactoring, and unit tests.

• Recurse Center is a "writers' retreat" for programmers of all experience levels.• I focused on writing ensemble machine learning algorithms from scratch, implementing new ML / computer vision methods from research papers, good Python programming practice, and some web development.

• Developed software for statistical simulations (using Langevin dynamics, Hamiltonian Monte Carlo & other techniques) to study shape transformations in cells (using C++), deployed these on distributed computing resources (using Bash / PBS), and analyzed data produced by these simulations (using Python).• Used git and Subversion for version control.• Built predictive models in Python based on first principles to explain collaborators’ experimental data.• Presented results in journal articles, posters, and 2 to 3 talks per year.

• Used modern quantitative experiments (FACS & microscopy + image analysis) to measure statistical variability in large populations of cells.• Cleaned, visualized, and presented these data to show these differences in observed cell phenotypes.• Discovered & shared insights about how the input-output of emergent signaling networks in yeast cells is fine-tuned.