Jack Daniel Reilly
Data scientist, data engineer, researcher, entrepreneur with a passion for mapping, geographical data, and machine learning at scale.
pdf • home • academic • linkedin • so • gh • jackdreilly@gmail.com • +1-916-768-1755
Education
- 2009-2014
- PhD, Civil Systems Engineering; UC Berkeley
- 2005-2009
- BSc, Civil Engineering; UCLA
Experience
Google Maps Machine Learning
Senior Software Engineer Tech Lead - 2014-2018 - Google - Mountain View, CA, USA
- Designed and deployed deep learning models at “Google scale” to extract mapping data from the world’s satellite imagery fused with geographical data, improving routing and travel time estimates for billions of users.
- Led, designed, and implemented Google’s first high-definition, self-updating road altitude dataset, instrumental in securing data deals with carmakers. This required data fusion, deep learning, model-driven optimization, visualization, and automated quality metrics to go from conception to launch.
- Miscellaneous:
- LSTM-based model to automatically detect fraudulent businesses on Google Maps
- Deep segmentation model to extract water body polygons from satelitte and infrared imagery.
Vianova Micromobility Analytics
Head of Engineering - 2020-2022 - Vianova - Paris, France
- Led, designed, implemented all backend tasks. Managed/mentored all backend/devops engineers.
- Implemented API to support complex, realtime geospatial analytics for scooter/bikeshare/careshare partners
- Implemented computation engine based on Snowflake unifying support for querying and batch-processing for millions of daily data points
- Led cultural initiatives around “tech talks” to bring more pride to people’s work and encourage sharing of technical hobbies.
- Design/Implemented innovative technology unifying data pipelines/plugins/sharing to turn our product into a next-gen platform for third-parties.
Cofounder and CTO - 2019-2020 - Paris, France
- Cofounded Taste to make a social-first food discovery platform.
- Designed, built and launched iOS, Android, Web MVP in 3 months
- Fully serverless architecture, modern, reactive front-end
Source control SaaS consultant - 2019
- Designed Kubernetes stack to scale infrastructure with increased userbase
- Migrated ad-hoc scheduling to tenant-isolated pods, with autoscaling
- Graphene+Prometheus monitoring solution
- Load balancing and networking across internal and external tools
- Designed and launched full-stack SaaS features, including DB, server, integration, front-end, testing, batch-processing
Google Maps Traffic Estimation
Software Engineering Intern - 2013 - Zurich, Switzerland
- I improved travel time estimations for directions on Google Maps by employing unsupervised Machine Learning clustering
- The discontinuity of travel times due traffic’s complexity makes naive approaches to travel time imputation inaccurate. I used unsupervised clustering techniques to identify discrete jams, and inferred per-jam travel times, leading to significant improvement in estimations, and enabling the tracking of jams as distinct semantic elements in Google’s backend.
PhD in Traffic Modeling and Optimization
At UC Berkeley, I developed model-based optimization and machine-learning methods for large-scale freeway networks. Based on distributed optimization ADMM techniques and discretized PDE networks, I developed a novel, decentralized model-based online controller for freeways, which enables the optimization of arbitrary objective functions. This allows city planners to devise complex goals for freeways which balance fairness and efficiency, or allow them to whimsically draw the CAL logo with car congestion.
Skills
by proficiency, over 11 total years, 6 years professional
Python • ML/Deep Learning • Java • Optimization • Mapping • Computer Vision • Big Data/Spark • C++ • Dart/Flutter • Git • Bash • Linux/OS-X • Java-/Typescript • Firebase/Firestore • Serverless • Protobufs • SQL • GCP • Docker • Tensorflow • Routing/Networks • Pandas • sklearn • Golang • Unit Testing • CI/CD • Prometheus • Kubernetes • AWS • GraphQL • Postgres/PostGIS • Reactive UI (React/Flutter) • Latex
Personal Interests
Traveled the world and volunteered for a year • Guitarist, Pianist • Acro Yoga • Runner • Jazz/Funk • Biking • Tennis • NBA • Swing Dancing
jackdreilly@gmail.com • +1 916 768 1755 Paris, France