Jonathan Frey
PhD Student At University Of Freiburg at University of Freiburg
Freiburg im Breisgau, Baden-Württemberg, Germany
Summary
🤩
RockstarJonathan Frey is a PhD student at the University of Freiburg with eight years of experience building high-performance numerical solvers and control software for robotics and autonomous driving. His contributions to prominent open-source projects—most notably optimizing the acados-based lateral MPC in comma.ai’s openpilot used on hundreds of cars—demonstrate a practical blend of control theory, low-level C/Cython optimization, and solver engineering. He has deep hands-on familiarity with interior-point and QP/QCQP solvers from work on acados, hpipm and the ASIPM project during an internship at Mitsubishi Electric Research Laboratories. Comfortable across research and production, Jonathan focuses on solver performance, parameter initialization, and tight low-level APIs that bridge algorithmic advances to embedded deployment. Based in Freiburg, he brings research rigor to real-world autonomy stacks and a track record of improving timing, residuals, and solver robustness in safety‑critical contexts.
8 years of coding experience
Github Skills (25)
robotics10
python10
c1110
controlling10
model-predictive-control10
mpc10
c1710
system10
control-theory10
controls10
sys10
control-flow10
controlled10
cython10
high-performance10
Programming languages (6)
Github contributions (5)
acados/acados
Feb 2018 - Jan 2023
Fast and embedded solvers for nonlinear optimal control
Role in this project:
Back-end Developer
Back-end DeveloperContributions:17 releases, 347 reviews, 2353 commits in 4 years 11 months
Contributions summary:Jonathan implemented and refactored functions within the acados template, focusing on improving the solver's functionality and efficiency. They added features such as parameter setting and improved the handling of algebraic variables in the cost module. The user's work included changes to the low-level C API, indicating a strong understanding of the solver's internal workings and the underlying mathematical optimization techniques. They also contributed to the addition of features for variable and parameter initialization.
nonlinearnonlinear-optimizationsolversoptimizationsolver
giaf/hpipm
Jun 2019 - Jan 2022
High-performance interior-point-method QP and QCQP solvers
Role in this project:Back-end Developer
Back-end DeveloperContributions:14 commits, 13 PRs, 3 comments in 2 years 7 months
Contributions summary:Jonathan primarily focused on refactoring and improving the performance of the `hpipm` library, which involves high-performance interior-point-method QP and QCQP solvers. Their work included optimizing core computational routines and fixing potential issues within the mathematical formulations. The changes involved modifications to C code, specifically in files related to the core algorithms for solving the optimization problems. Additionally, there were adjustments made to the Python interface for correct use with the MATLAB integration, along with a bug fix for handling empty-shape inputs.
solversmethodoptimizationmultiobjective-optimizationconstrained-optimization
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.