Humboldt-Universität zu Berlin - High Dimensional Nonstationary Time Series

Bruno Spilak



bruno.spilak [at] hu-berlin [dot] de




Office hours:

Dorotheenstr. 1

Upon agreement

Postal address:

Ladislaus von Bortkiewicz Chair of Statistics
School of Business and Economics
Humboldt-Universität zu Berlin
Unter den Linden 6
10099 Berlin, Germany



2018 - 2023 PhD student in statistics, Humboldt-Universität zu Berlin  
2016 - 2018 Master of Science in Statistics, Humboldt-Universität zu Berlin and ENSAI (joint degree)
2014 - 2016 Master of Science in Statistics, ENSAI, Rennes, France
2014 - 2015 Bachelor of Science in Mathematics, Université de Rennes 1 and ENSAI (joint degree), Rennes, France
2012 - 2014

Post-secondary preparatory classes, Khâgne BL (Classes préparatoires aux grandes écoles, Khâgne BL – CPGE) - Lycée Henri IV, Paris, France


Research Interest

  • Nonparametric Statistics

  • Machine learning

  • Reinforcement learning
  • Quantitative Finance
  • NLP





  • Summer camp, Buckow, 2018: Deep Neural Networks for Cryptocurrencies Price prediction

  • Haindorf Seminar, Haindorf, 2019: Deep Reinforcement Learning
  • STAT of ML, Prague, Czech Republic, 2020: Tail-risk protection: Machine Learning meets modern Econometrics
  • 29th Annual Conference on PBFEAM, Rutgers University, New Brunswick, NJ, 2021: Tail-risk protection: Machine Learning meets modern Econometrics
  • ML approaches Finance and Management, HU Berlin, 2022: Does Non-linear Factorization of Financial Returns Help Build Better Portfolio?


Work in Progress

  • ReLU networks and nonlinear PCA
  • News sentiment analysis and portfolio
  • DAI, NFT prices: NFT images style embedding





Presentations, courses, talks, Github and Quantlets logo_Quantnet