Matt Matolcsi

About Matt Matolcsi

Interested in solving foggy problems that need someone to understand the context and the data (and that deliver value!) Also interested in statistical, machine learning, and optimization methods of all kinds, and mentoring team members. Python, SQL, Linux. Optimization methods, including Gaussian Process-based Bayesian optimization approaches (BoTorch). Regressors, classifiers, time series anomaly detection. Visualizations, analyses, reporting. Frequentist and Bayesian methods, including Hidden Markov Models. CI/CD, basic DevOps. Docker, Docker Swarm. Git, GitHub Actions.

Experience

1. Senior Mle & Data Scientist

   Viasat

   Apr 2021 — Present · Carlsbad, CA, US

   Co-leading the development of an in-house dev tool bringing a novel approach to RF R&D. I focus on algorithm development, project planning, customer success, workflows, visualizations, and documentation.* Co-researcher on a next-gen satellite capability; it\'s a tough problem.* Led/Co-led the data science effort to understand a root cause of equipment failure and develop predictive models to get advance warning. This suite of models has been continuously in production since late 2021 and covers the equipment responsible for about 1/3rd of Viasat\'s revenue.* Advocating for adoption of new tools and capabilities among the team* Mentoring and providing feedback to colleagues

Education

• UCLA

  Bachelor of Science (B.S.), Physics

  1999 — 2004

• Purdue University

  Doctor of Philosophy (Ph.D.), Physics

  2007 — 2014

• Purdue University

  Certificate, Applied Statistics

  2013 — 2014

Skills

• Simulations
• C
• Programming
• Optimization
• Machine Learning
• Python
• Latex
• Markov Chain Monte Carlo
• Matlab
• Sas
• Sql
• Linux
• Statistics
• C++
• Data Analysis
• Scala
• R
• Hdf5
• Algorithms
• Hungarian
• Analysis
• Physics
• Mathematical Modeling
• Experimentation
• German
• Data Mining
• Science
• Apache Spark
• Research
• Numerical Analysis
• Modeling