Solving this problem: Data scientists need financial data formatted in a specific way and continuously updated in order to build models. Further, those models need rigorous backtesting compared with Monte Carlo simulations to assess true viability

• Took Sift from startup to Inc 5000, Deloitte Tech Fast 500, Top Powerlist for AppsFlyer• Data science increased topline revenue 85%, bottomline by 127% (higher rev, much higher margins)• Headed data science team: hiring, development road map, recruiting, planning• Created Sift’s primary real-time, multidimensional bidding algorithm and backend data framework• Created robust experiment framework and reports that quickly identify pockets of significance• Created several realtime report tools to give data and business insights into performance and traffic changes

DT's bread and butter is delivering the right apps to the right people. To power that decision engine, I drove several data science initiatives: • Partnering with the revenue team to produce a campaign optimization tool that predicts volume based on changes in bid rate. • Collaborative filtering algorithms utilizing the Scala flavor of Spark. • Comprehensive reports on the statistical significance of all user features available to our bidder decisioning algorithm.

Helped Quandl through their pre-acquisition stage; UA was of paramount importance.* Clearbit API integration for automatic lead discovery* A/B tests and cohort analysis for messaging and UA efforts* Exporting our API to DuckDuckGo for Quandl-branded search results* Dozens of Github public repo updates in at least 6 different languages to better utilize the Quandl API (better code integration; more subscribers!)* On-site presentations/evangelizing at FinTech Sandbox, Quantopian hackathons and Tufts

Drupal /PHP/MySQL/SASS/UI/UXdevelopment for client portals and cross-platform tablet apps. The real pleasure of this position was working with delightfully talented people who are passionate about their projects. I performed interface and data visualization work, so a primary challenge has been reducing these hyper-dimensional data sets to their essential, most digestible elements. Simplicity is often quite difficult, but highly worth the effort. It takes planning, foresight and a certain degree of courage to *not* display certain data elements. However, thoughtfully choosing how to distill and reduce... that is a large part of the craft of clarity.

An amazing position -- this introduced me to the exciting field of proteomics. Here I was able to apply my math background along with my programming skills to create top-notch bioinformatics software that helps researchers find disease-specific (especially cancer) forms of proteins. The confers a tremendous amount of potential to identify novel methods for diagnosis (biomarkers) and drug targets that *only* affect unhealthy protein pathways.

Geneffects produces custom software, primarily for biotech, fintech and analytical marketing. Geneffects' flagship product, Peppy, is proteogenomic protein identification software tuned to identify novel (non-canonical) proteins in tandem mass spectrometry data. Applying this software to data from cancer samples can identify novel proteins that are focused, highly promising candidates for biomarkers for early detection of the disease with results published in Nature and the Journal of Proteome Research

Lead developer for internal content management system written in Java (J2EE). This job was great in that I it played off of my front-end and back-end development skills. There were all kinds of users: Principal investigators, investigators, clients, etc., so permissions granularity and intelligent data sharing were key priorities. In general, though, this was a fun project because, at its heart, it was about creating a vehicle for fostering collaboration.