Jonathan Dumm Email & Phone Number

Stockholm, Stockholm County, Sweden

Stockholm, Sweden

Using Big Data, SQL (Hive/Spark), statistical (A/B) tests, machine learning, and mathematical modeling to improve caller ID and spam protection for smart phones.

* Led the muon analysis working group for 2.5 years, guiding efforts in software development (python, C++) and machine learning for scientific advancement of the IceCube Collaboration. * Served on the publication committee for 2 years, reviewing and guiding collaborative authors. * Developed a new maximum likelihood analysis to identify diffusive features of the sky in large data sets, used to look for dark matter and cosmic-ray interactions in the Milky Way. * Implemented real-time followup program, sending alerts from the South Pole to the VERITAS telescopes in Arizona with sub-minute latency, including automated telescope repointing. * Helped develop a simulation of protons propagating in the universe, used to infer their origins and model emission from blazars. * Used scikit-learn's machine learning suite (BDTs and neural networks) to improve energy estimation of muons moving through the Antarctic ice sheet. * Mentored several graduate and undergraduate students.

Greater Minneapolis-St. Paul Area

* Served as Offline Analysis Working Group coordinator and lead analysis software developer (C++) for 3.5 years.* Developed the VERITAS Blazar Pipeline for high quality next-day automated analysis with an accompanying web page.* Built a custom nightly build (CI) for the project with a web interface for noting and tracking changes. * Developed advanced pattern recognition (image analysis) software, leading to two new gamma-ray blazar detections.* Optimized background rejection using machine learning techniques (Boosted Decision Trees).* Wrote time allocation committee proposals for the blazar discovery program.* Led multi-day “software boot camp” training sessions for both my team as well as the entire field. * Mentored several graduate and undergraduate students.

Madison, Wisconsin Area

* Developed maximum likelihood statistical models for identifying potential sources of astrophysical neutrino sources including confidence level construction.* Designed “level 1” high efficiency, robust filters to select useful events for transmission from the South Pole over satellite for physics analysis.* Benchmarked and refined muon track fitting techniques to produce a "standard processing” used by nearly all 300 collaborators.* Designed Support Vector Machine (SVM) and Boosted Decision Tree (BDT)/Random Forest event classifications for extracting 1 part in 1 billion signal events from backgrounds.* Developed and benchmarked Monte Carlo simulations.* Traveled to the South Pole during the Austral summer of 2008–2009 to perform final acceptance tests on hardware and to deploy instruments kilometers deep in South Pole ice.* Mentored both undergraduate and graduate students as well as ran collaboration-wide “software boot camp” training sessions.

Doctor Of Philosophy (Phd), Physics

Bachelor'S Degree, Physics