Florian Wolf

Hi, I am a PhD student in Applied and Computational Mathematics (ACM) at Caltech, working with Prof. Andrew Stuart and supported by the Kortschak Scholars Program Fellowship.

Industry. This summer, I will join the Amazon AWS AI Labs in Santa Clara, working with Danielle Maddix Robinson and Yuyang (Bernie) Wang . Before joining Caltech, I was an intern at Amazon Robotics in the Vulcan Pick Team. Between Bachelor's and Master's, I was working at Mercedes-Benz as a research and development engineer intern.

Education. I hold Master's degrees in Mathematics and Computational Engineering from TU Darmstadt. For my thesis, I worked on Hidden Convex Optimization with Functional Constraints in the Optimization and Decision Intelligence Group at ETH Zürich, supervised by Prof. Niao He. During my studies, I was honored to be a fellow of the German Academic Scholarship Foundation ("Studienstiftung") as well as the Swiss National Centre of Competence in Research (NCCR) Automation, and I was supported by the Prof. Dr.-Ing. Erich Müller-Stiftung and the Almetes Stiftung.

My research interests include

• Optimization and Reinforcement Learning
• (Multimodal) Representation Learning

Email  /  Scholar  /  Github  /  LinkedIn

Thesis / Project supervision:

If you're passionate about research and interested in working with me, I'm always happy to hear from motivated students. If you'd like to get involved, please reach out with a summary of your research interests, along with your up-to-date CV and academic transcript.

While maximum-entropy exploration is central to RL, real-world applications are often hindered by safety constraints and a lack of additive structure. We propose Policy Gradient Penalty (PGP), a single-loop method that enforces general occupancy-measure constraints via quadratic penalty regularization. By leveraging hidden convexity, we provide the first global last-iterate convergence guarantees for this non-convex setting.

First algorithms with provable global guarantees for constrained non-convex optimization via hidden convexity. Our methods bypass constraint qualifications, handle hidden convex equality constraints, and work directly with gradient oracles in the non-convex space. Applications include safe control and reinforcement learning, where we establish global optimality and oracle complexities matching unconstrained hidden convex optimization.

AE+SINDy-C: a data-efficient, interpretable, and scalable Dyna-style Model-Based RL framework for PDE control, combining SINDy-C with autoencoders for dimensionality reduction. Applied to the 1D Burgers and 2D Navier-Stokes equations, the method enables fast rollouts, reduces environment interactions by up to 10x, and yields an interpretable latent dynamics model, outperforming a model-free baseline.

Bayesian spatio-temporal product partition model to cluster PM2.5 air quality data from multiple monitoring stations in Northern Italy, capturing both spatial and temporal patterns. The model outperforms a spatial-only baseline in predictive performance, offering smoother and more insightful pollution trend analysis.

Automated curriculum generation with massively parallel RL learns a spherical pendulum tracking controller, leveraging the task's non-Euclidean structure for faster convergence, higher performance, and successful sim-to-real transfer.

The website is based on the code from Jon Barron