Who reviews and tests food tracking apps at Food Tracker Compass?
Food Tracker Compass is run by a small, fully-independent group of AI researchers, ML engineers, nutrition scientists and UX leads. Between us, we have built or evaluated multimodal vision systems at Google, Meta, AWS, Spotify, Stripe, Snap, Cohere and Duolingo, and worked with food-composition data at the EFSA, Open Food Facts and USDA. The benchmark is what we wished existed when we left those jobs.
Dr. Elena Marquez
PhD Computer Vision, ETH Zürich · Zürich, Switzerland
Designs the recognition-accuracy half of our benchmark and signs off on every protocol version. Twelve years building and evaluating multimodal vision systems, with a focus on long-tail food-image recognition.
Read full profile →
Jin Kobayashi
MS Computer Science, Carnegie Mellon · Tokyo, Japan
Runs the day-to-day evaluation lab, owns the comparative reviews, and writes the test harness that drives every app through the protocol. Eight years shipping production ML on mobile.
Read full profile →
Dr. Sara Owusu
PhD Machine Learning, University of Toronto / Vector Institute · Toronto, Canada
Owns reproducibility for the benchmark — double-grading, statistical sign-off, and the publication review for every individual app review. Authors our methodology change log.
Read full profile →
Daniel Reinhart
MS Data Engineering, TU München · Munich, Germany
Builds and maintains the test pipeline — capture, hashing, storage, scoring, and the dashboards that let us see drift over time. Operates the lab infrastructure end to end.
Read full profile →
Dr. Amelia Novak
PhD Food Science, Wageningen University · Amsterdam, Netherlands
Validates the food and nutrient data layer of every benchmarked app against USDA FoodData Central and Open Food Facts. Owns the Database and Nutrients scoring categories.
Read full profile →
Marcus Hale
MSc Human-Computer Interaction, University College London · London, United Kingdom
Owns the Ease of Use category of the rubric. Runs the participant task-completion timing and accessibility audits, and writes the user-facing review prose.
Read full profile →How does the Food Tracker Compass team divide up the work?
Elena owns the recognition-accuracy half of the benchmark and signs off on every protocol version. Jin runs the day-to-day testing lab and writes the comparative reviews. Sara owns reproducibility, the double-grading process, and the publication review for every individual app review. Daniel builds and maintains the data pipeline — capture, hashing, storage and scoring. Amelia validates the food and nutrient data layer of every app against USDA FoodData Central and Open Food Facts. Marcus designs the participant tasks behind every Ease of Use score and writes most of the narrative review prose. Every protocol change is reviewed by all six of us, and the protocol version number lives in git history.
For methodology details, see the methodology page. For our editorial standards and funding model, see the about page.