Ben Nelson Email & Phone Number

Redwood City, CA, US

• Full stack product analytics that supports bioimaging technology and programs
• ETLing data for AI models
• Data science intern manager

Redwood City, CA, US

Product analytics for a biomedical research discovery app

San Francisco, California, US

Palo Alto, California, US

• Data scientist for Wealthfront's free financial planning product.
• Designed metrics for product area
• Conducted in-app A/B experiments to improve onboarding and account funding conversion (R, tidyverse)
• Performed demographic analyses to predict and learn about user engagement with the Wealthfront app (R, tidyverse)
• Wrote several data pipelines, enabling Wealthfront to monitor all email-related AB experiments, affiliate performance, and different account funding metrics (git, Scala, Spark)
• Built and maintained internal shiny dashboards for financial planning, blog content, and marketing (R, shiny)

San Francisco, CA, US

• Built WhirlWind, a web app that recommends time-efficient walking tours of San Francisco attractions
• Scraped, processed, and engineered features from Tripadvisor reviews
• Developed a pipeline using k-means clustering + Dijkstra's algorithm to generate "highly-rated" walking routes and validated my recommendations on hundreds of simulated cities.
• Deployed whirlwind.site with an interactive map using Flask, AWS, Google Maps API, and Folium

Evanston, IL, US

• Led original interdisciplinary research projects
• Developed hierarchical Bayesian models using Stan to infer multiple populations of exoplanets from real datasets (mc-stan.org)
• Collaborated with civil engineers to apply a network-based frequency analysis to uncover more realistic features in exoplanet distributions (csun.uic.edu/codes/NFA.html)
• Designed and organized a worldwide data challenge for the exoplanet and astrostatistics communities to quantify the "evidence" for n-planets in time series data (github.com/EPRV3EvidenceChallenge/Inputs/)Collaborated.
• Developed an efficient pipeline to compute quantitative evidences of competing orbital models (i.e., Bayes factors) for thousands of real and synthetic Doppler time series dataAdvised one graduate and two undergraduate.
• http://adsabs.harvard.edu/abs/2018ApJ...867...75DInvited to and participated in grant selection committees at NU and NASAFounded and organized efforts to enhance data science activities at NU

University Park, PA, US

• Led two analyses of dynamically complex multi-planet systems
• Estimated the planet masses and orbital parameters using our homebrewed differential evolution Markov chain Monte Carlo (C++/CUDA) applied to their star's Doppler observations
• Determined if each orbital model was dynamically stable using the MERCURY integrator (Fortran)Wrote an importance sampling algorithm (Python, numpy) to compute the Bayesian evidence for n-planets in data.

Gainesville, Florida, US

• Developed code in C++ for a differential evolution Markov chain Monte Carlo applied to exoplanet observations and wrote a CUDA version that attained ~5x speedup over the OpenMP versionRegularly sent batch jobs to UF's.
• teaching assistant for intro astronomy class (4 semesters)
• guest lecturer for undergraduate/graduate level exoplanets class
• guest author for astrobites.comDepartment webmaster for 1 year

Doctor Of Philosophy (Phd), Astronomy And Astrophysics

Phd Candidate, Astronomy

Master Of Science (M.Sc.), Astronomy

Bachelor Of Science (Bs), Physics (Astrophysics)