About Me

I recently graduated from the Erasmus University Rotterdam obtaining a Master of Science (M.Sc.) degree in Econometrics and Management Science. The master’s track I have undertaken was Operations Research & Quantitative Logistics. For my master’s thesis, I was advised by Michel van de Velden. For the related research, I visited Berkeley Artificial Intelligence Research lab (BAIR), where I was advised by Pieter Abbeel. I have a Bachelor of Science (B.Sc.) degree in Industrial and Applied Mathematics from Delft University of Technology. For my bachelor’s thesis, I was advised by both Marco Loog and Frank van der Meulen.

See my full resume for more details.


03/2020 - Present

Machine Learning Research
UCL Centre for Artificial Intelligence

Working with Thomas Bird and David Barber from UCL on more memory efficient deep generative models.

11/2019 - Present

Migration Project

Working with a large group of students to migrate merchant information from sheer documents into a data-driven system.

09/2018 - 06/2019

Visiting Student Researcher
Berkeley Artificial Intelligence Research (BAIR)

Did research at the junction of deep unsupervised learning and information theory, advised by Pieter Abbeel and Jonathan Ho. Designed a novel compression algorithm called “Bit-Swap” that uses a recursive extension of bits-back coding to exploit hierarchical VAE’s. This method is able to outperform many established lossless compression techniques on images.

12/2015 – 08/2018

Functional Designer and Consultant

Responsible for designing, maintaining and (on request) modifying construction process management systems (part of BIM) for various contractors. Advising contractors about project management in Relatics and designing automatic reports. Advanced skills in the online software Relatics.


Friso H. Kingma

Improving Data Compression Based On Deep Learning

M.Sc. thesis, Erasmus University Rotterdam
The Netherlands, 2019
[thesis] [citation]

Friso H. Kingma, Pieter Abbeel, Jonathan Ho

Bit-Swap: Recursive Bits-Back Coding for Lossless Compression with Hierarchical Latent Variables

Proceedings of the International Conference on Machine Learning (ICML)
Long Beach, USA, 2019
[paper] [citation] [talk] [slides] [poster] [code] [blogpost] [berkeley lectures ’19] [berkeley lectures ’20]

Friso H. Kingma

BagDrop: Computationally Feasible Approximate Bagging with Neural Networks

B.Sc. thesis, Delft University of Technology
The Netherlands, 2017
[thesis] [citation]


2017 - 2019

Master of Science (M.Sc.)

Econometrics and Management Science
Operations Research and
Quantitative Logistics

Erasmus University Rotterdam

2013 - 2017

Bachelor of Science (B.Sc.)

Industrial and Applied Mathematics
Delft University of Technology

Prof. Pieter Abbeel


Prof. Marco Loog


Prof. Michel van de Velden