Business intelligence supporting blockchain organizations and developers. • Geotagged cryptocurrency addresses with a fusion of on-chain and off-chain signals.

Anti-money-laundering of cryptocurrency networks for banks and government clients.•Categorized & labeled 30M crypto addresses with unsupervised machine learning.•Identified key metric predicting wallet “coldness”. Forecasted Ethereum transactions.•Designed & prototyped C.T.’s first Ethereum visualization. Clarified complex address-address relationships.

Project management of cloud-based machine learning pipelines. Experimental Social Computing. Teaching.•Led technical development of a civil unrest forecasting engine streaming open source data.•Orchestrated team: 7 SWE's, 30 machine learning researchers, 10 social scientists.•Achieved 80% precision, 70% recall predicting events later verified in journals of record.•Secured high-value contract delivering civil unrest warnings to U.S. intelligence community.•Introduced agile methodology so students could build modular, reusable research software.•Tracked cross-platform content sharing with novel “social honeypots” and snowball sampling.•Taught “Introduction to Data Analytics and Visualization.” (70 students. Managed 4-6 teaching assistants.)

Cofounded digital advertising company serving mobile ads based on user’s TV ad exposure.•Designed Naive Bayes machine learning system predicting TV viewership of mobile users.•Inferred viewership & demographics of 100M mobile users with app history & set top box data.•Scrum master for a cross-functional team of data scientists, software developers, and dev ops.•Presented to clients & investors on the value proposition of a synergistic TV & mobile advertising strategy.•Achieved a $1M per year run rate within the first year of operation.•Campaigns ranged: 100k-1M in reach, and $50k-$1M in ad buy. •Achieved 3X higher click-through rate than competitors serving identical mobile creative content.

Designed internal platforms analyzing sales revenue, sales compensation, and promotion.•Exploited natural experiments created by non-uniform application of incentive programs. Overturned prior beliefs about effects of overweighting comp on upselling new products.•“Myth-busted” the relationship between PERF and salesperson performance.Designed internal platforms analyzing sales revenue, sales compensation, and promotion.•Conducted ethnographic-style interviews validating unexpected results; established ground-truth of how salespeople perceived incentives. Designed dashboard around possible cognitive bias of comp designers.

Machine learning for dynamic pricing of live-event tickets. (SF Giants, Philadelphia Flyers, and Live Nation.)•Mined historical data for relationships between exogenous events, prices, and sales.•Developed probabilistic, multivariate, time-dependent models of ticket revenue.•Automated daily delivery of pricing recommendations to box offices.•Demonstrated increased revenue for SF Giants through an A/B test between outfield sections.

Built the second radio-frequency scanning tunneling microscope under ultra-high-vacuum. RF circuit and signal design, as well as design of a novel RF-STM tip-exchange system.