Torc Robotics builds autonomous driving software for Class 8 heavy trucks, acquired by Daimler in 2019 to accelerate commercialization. The tech stack—C++, CUDA, PyTorch, TensorRT, LIDAR perception—reflects a mature autonomous systems engineering org heavily invested in inference optimization and simulation. Active projects span offline perception model development, BEV (bird's-eye-view) models, and closed-loop validation, while documented pain points (inference on embedded hardware, domain data gaps, sensor data scale) point to a company solving real production constraints rather than research problems.
Tech Stack 52 technologies
What Torc Robotics Is Building
◆Challenges
- Optimizing inference performance on embedded hardware
- Improving perception accuracy
- Domain data gap
- Massive sensor data processing
- Scaling perception simulation
- Maintaining high uptime
- Reliable software delivery to autonomous vehicles
- Model maturity
- Scaling rl behavior models
- Collecting large behavior datasets
▲Active Projects
- Road & lane detection
- Neural rendering framework for perception simulation
- Ml development pipeline
- Offline perception model development
- Bev model development roadmap
- Offline perception annotation pipeline
- Data ingestion and governance pipeline
- Closed-loop simulation and on-road validation
- Ml framework improvement
- Large-scale training workflow
Hiring Activity
Steady55 roles · 15 in 30d
Department
Seniority
Notable leadership hires: Chief Financial Officer
Company intelligence
