Data, Models, Optimization.

About Me

I am a Machine Learning expert and engineer with more than ten years of experience in various fields. I  have worked for Airbus as well as for several Formula 1 teams (McLaren-Mercedes, Ferrari and Red Bull). I currently work for the New Zealand America's Cup team and offer consulting services in data science and engineering (see section below).

I have combined my applied work with PhD research at the University of Cambridge. This has given me the exceptional opportunity to work with leading academics in the filed of Machine Learning and to connect my real-world experience with exciting new computational techniques for data analysis and predictive modeling.

More about me on my LinkedIn profile.

More about Machine Learning in this excellent article by McKinsey.

Consulting Services

I offer consulting services in data science and engineering. If you are interested in any of the following topics please feel free to email me [] or contact me on LinkedIn to explore a potential collaboration.

  * Predictive Modeling: Do you want to gain insight and make predictions based on data?

  * Optimization: Do you want to use predictive models to optimize revenue, lap time or any other goal? How do you optimize multiple conflicting goals?

  * Design of Experiments: If you have the chance to collect new data, what data will improve the predictive models the most? Which experiment should you run next in order to optimize your goals?

  * Uncertainty Quantification: How confident are you in your estimates? Is your current data informative enough to make confident predictions?

  * Dynamical Systems and Control: Do you want to model, identify, simulate or analyze the stability of linear and nonlinear dynamical systems?



 Why did I do a PhD in Machine Learning?

 Dealing with Uncertainty in Engineering

 Why Should All Engineers Master Statistics?

PhD Thesis

My PhD thesis focused on learning models of nonlinear dynamical systems based on measured data. Those models rely on Gaussian processes and can provide probabilistic descriptions of uncertainty. My work resulted in new insights on the mathematical description of the models and the development of novel learning algorithms based on those insights.

Bayesian Time Series Learning with Gaussian Processes,
University of Cambridge, PhD Thesis, 2015.
[pdf] [bib]


Variational Gaussian Process State-Space Models,
R. Frigola, Y. Chen and C. E. Rasmussen.
Advances in Neural Information Processing Systems (NIPS), 2014.
[pdf] [bib]

Identification of Gaussian Process State-Space Models with Particle Stochastic Approximation EM,
R. Frigola, F. Lindsten, T. B. Schön and C. E. Rasmussen.
19th World Congress of the International Federation of Automatic Control (IFAC), Cape Town, South Africa, 2014.
[pdf] [bib]

Bayesian Inference and Learning in Gaussian Process State-Space Models with Particle MCMC,
R. Frigola, F. Lindsten, T. B. Schön and C. E. Rasmussen.
Advances in Neural Information Processing Systems (NIPS), 2013.
[pdf] [bib]

Integrated Pre-Processing for Bayesian Nonlinear System Identification with Gaussian Processes,
R. Frigola and C. E. Rasmussen.
52nd IEEE Conference on Decision and Control (CDC), Florence, Italy, 2013.
[pdf] [slides] [bib] [code]

A Wrench-Sensitive Touch Pad Based on a Parallel Structure,
R. Frigola, L. Ros, F. Roure and F. Thomas.
IEEE International Conference on Robotics and Automation (ICRA), Pasadena (California), USA, 2008.
[pdf] [bib]


Engineering in the Age of Machine Learning,
Meet Up @ Kernel Analytics, Barcelona, 30 October 2015.

Gaussian Process Models for Nonlinear Time Series (with Carl E. Rasmussen),
Tutorial, Cambridge, 16 April 2015.

Learning Dynamical Systems with Gaussian Processes,
Research Talk, Cambridge, 24 February 2014.

Probabilistic Models for Big Data (with Alex Davies),
Tutorial, Cambridge, 13 February 2014.

Integrated Pre-Processing for Bayesian Nonlinear System Identification with Gaussian Processes,
52nd IEEE Conference in Decision and Control, Florence, 12 December 2013.

Tutorial on Bayesian Nonparametric Nonlinear System Identification (with Andrew McHutchon),
ERNSI Workshop, Nancy, 25 September 2013.

Bayesian Nonparametric Nonlinear System Identification,
Reglerteknik Monday Meeting, Linköping University, 10 June 2013.

Learning to Control: State Estimation,
Research Talk, Cambridge, 30 April 2012.

Statistical Inference for Engineers,
Seminar, Maranello, 19 March 2012.

An Overview of Control Theory,
Tutorial, Cambridge, 12 January 2012.


Cambridge University Engineering Department
Trumpington Street
Cambridge, CB2 1PZ
United Kingdom-

View Roger  Frigola's profile on LinkedIn

Cambridge logo