MSc Thesis: Quality-Ranked Training on Easy Reasoning Problems
Going beyond the RLVR paradigm for training math reasoning models. At SRI, ETH Zurich with Jasper Dekoninck and Prof. Martin Vechev.
Hanno Hiss
I'm currently building something new.
Previously, I studied at 
ETH
Zurich and 
MIT, graduating with an MSc in Data Science (minor in Robotics). My thesis was advised
by Prof. Martin
Vechev at 
SRI. I've also worked at 
QuantCo on statistical models for insurance companies. Before
that I joined 
Swissloop to build vehicle control and sensors for the future of transportation.
I've long been drawn to statistical learning and especially reinforcement learning, and I care most about building
things that matter.
Outside of work, I like running marathons and cooking.