▪ Analyzed 2+ million farmer profiles and 100’s millions engagement data points for PxD’s India-based agricultural advisory service, to monitor operations and derive insights on farmer pick-up, listening, and other engagement metrics.▪ Built rice-yield prediction model using remote sensing to make impact evaluations more cost effective: from data prep (e.g. processing PlanetScope and Sentinel-2 RS data in GEE and python into VI-features), to data QA (e.g. developing a custom Google Earth Engine (GEE) app to perform QA and correct 14k+ collected plot boundaries), to model training.▪ Developed Django-based AWS-backed website (front & back-end) to enabled 100+ call center staff to continue onboarding farmers from home (pandemic) using custom SurveyJS integration, and to host an A/B-tested self-enrollment webform for farmers, reducing the marginal cost of onboarding farmers to ~zero.▪ Initiated and executed data cleaning and standardization processes (e.g. from dta-files to postgres; standardizing farmer location data using fuzzy matching and google maps API; automating data pipelines from 3rd parties to internal systems).▪ Supported operations and research (e.g. sampling, survey design, field ops, map creation, data security, analysis)▪ Maintained PxD’s global AWS-backed systems (Mattermost, ODK central) and piloted a WhatsApp chatbot.

▪ Developed Root Capital’s current Probability of Default (PD) risk model, resulting in reduced portfolio risk and a projected annual loss-reduction of ~$500,000 on ~$65MM of investments. (In python using LASSO Logit model)▪ Data wrangling and cross-departmental aggregation to drive novel research related to impact, growth, and performance.▪ Analyze borrower’s growth paths and access to finance: get historical trends, find drivers of growth, and generate insights.▪ Reworked the quantitative and analytical underpinnings of Root Capital’s Efficient Impact Frontier (EIF) framework.

Laterite is Rwanda's most reputable data, research, and advisory firm, offering full life-cycle projects: from design to data collection, analysis, and policy advice. Clients range from NGO’s and firms to World Bank and government.As a consultant to Laterite, I developed predictive models of repetition and dropout for the Ministry of Education to predict and preempt individual pupil repetition and dropout.

▪ Developed sampling and survey strategy to meet stratification requirements, maintain i.i.d., and increase efficiency.▪ Performed analysis, research, and writing on education and teacher training to optimize Mastercard Foundation’s funding.▪ Created a model for the Ministry of Education to predict and preempt pupil repetition and dropout.

▪ Analyzed impact of shocks on financial flows, financial device use, and remittances of Kenyan households by performing econometric analysis (Stata) on 11-month panel data of financial flows and device use by 300 households.

As part of the Belgian development agency's €8MM project in Uganda that sought to improve institutional capacity at the Ministry of Health, I:▪ Established the Ministry of Health’s eLearning program for continued professional development of health workers.▪ Presented at the 2014 eLearning Africa conference, to 700+ participants from 70 countries.

In partnership with 12 companies, in this 2-year project to increase private sector adoption of OPC UA, a standard protocol in data and web-based industrial communication, I:▪ Developed C# implementations of OPC UA client/server software; supported project partners’ software development.▪ Organized presentations, workshops, and a one day conference with 90+ international attendees.