Portrait of Alberto Gennaro
Incoming Senior Associate · J.P. Morgan Chase — ML Centre of Excellence (Time Series & RL), NYC
Ph.D., UC Berkeley IEOR

Hi, I’m Alberto. I’m about to join J.P. Morgan’s ML Centre of Excellence in New York as a Senior Associate in Machine Learning, on the Time Series and Reinforcement Learning team.

Before that, I did my Ph.D. at UC Berkeley IEOR, advised by Prof. Thibaut Mastrolia. My research focused on stochastic optimal control under time and model uncertainty, with applications in financial engineering.

Earlier I did my Bachelor’s and M.Sc. in Mathematical Engineering at Politecnico di Torino, alongside a second M.Sc. in Statistics and Applied Math at Collegio Carlo Alberto. My thesis was advised by Prof. Brandimarte and Prof. Fadda, on tractable approximations of large-scale multi-stage stochastic programs via linear decision rules.

Outside of work you’ll find me on a tennis court or a golf course, or trying to learn something new. If you work in RL or time series, I’d love to hear from you at alberto (dot) gennaro (at) berkeley (dot) edu.

Time Series Reinforcement Learning Stochastic Control ML / Deep Learning Numerical PDEs
CV available upon request — email me.

News

Recent activity
Soon Joining J.P. Morgan Chase — ML Centre of Excellence, on the Time Series & Reinforcement Learning team.
May 2026 Ph.D. defended at UC Berkeley IEOR. Also: invited talk at Bachelier FS26 (Bologna).
2026 Delegated portfolio management with random default accepted for publication in Mathematical Finance.
Oct 2025 2nd place at the INFORMS RAS Problem Solving Competition (with team MathConvoy), INFORMS Annual Meeting.
Aug 2025 Wrapped up Applied Scientist internship at Amazon SCOT (Bellevue, WA).
Jul 2025 Best Paper Finalist (top 6) — SIAM Conference on Financial Mathematics (Miami).
Jun 2025 Preprint posted on arXiv: 2BSDE with uncertain horizon.

Research focus

All projects

What I work on

Sequential decision-making under uncertainty: time-series forecasting, reinforcement learning, stochastic control with uncertain horizon, and the numerical methods that make them computable.

  • Time-series forecasting and policy learning from sequential data.
  • Reinforcement learning for control — including signature-based methods.
  • Stochastic control with uncertain horizons; PINN solvers for HJB-type PDEs.

Selected papers

2BSDE with uncertain horizon and application to stochastic control in erratic environments
Preprint · arXiv:2506.15037 · 2025
Delegated portfolio management with random default
To appear in Mathematical Finance · arXiv:2410.13103 · 2024

See all publications →

Highlights

Jul 2025

Best Paper Finalist (top 6) — SIAM Conference on Financial Mathematics

Oct 2025

INFORMS RAS Problem Solving Competition — 2nd place

2025

Outstanding Graduate Student Instructor Award — UC Berkeley

Get in touch

Email

(my name) dot (my last name) (at) (berkeley dot edu)

Best for collaboration, research questions, and CV requests.