![Banseog [ physical ai ]](/assets/brand/banseog-physical-ai-on-dark.svg?v=20261007){kind=small}
Banseog note · 2026-10-25
The software center building Hyundai Motor Group's own autonomous driving AI
Expansion — The production schedule for in-house autonomous driving AI is set, and the company is recruiting experienced talent at scale to match it
- According to Korea's National Pension Service public workplace data, there were 773 people on the pension register in August 2026, 78 more (11.2%) than in September 2025. In that period 297 people joined and 219 left.
- At the Hyundai Motor Group autonomous driving media day held at 42dot's Pangyo headquarters on September 11, 2026, the company said it operates more than 40 IONIQ 5-based data collection vehicles and more than 20 IONIQ 6-based SDV test vehicles, collecting over 80 hours of driving data per vehicle each week.
Decisions within a year
- Production of in-house autonomous driving AI: The company plans to launch Level 2++ vehicles based on 'Atria AI', co-developed with the Hyundai Motor and Kia AVP Division, in the second half of 2029 (preceded by NVIDIA-based production in 2028).
- Start of VLA autonomous driving development: In September 2026 it began developing VLA-based autonomous driving that combines visual perception, language reasoning and vehicle action in a single model, and said it will fully launch the development process, including real-vehicle testing, from the end of this year into early next year.
- Large-scale experienced hiring From July 23 to August 5, 2026, it recruited up to about 70 experienced staff for more than 30 roles across 7 divisions, and plans to hire about 200 new people by the end of this year. In September and October it externally recruited a robotics middleware expert as principal research fellow at its Silicon Valley office, along with a VLA team lead and a voice AI model group head.
People they will look for
- Autonomous driving deep learning and VLA engineer — the company needs people to carry Atria AI and the VLA model through real-vehicle testing and into 2029 production
- Driving data and training infrastructure engineer — the data flywheel must keep running, selecting from the data that dozens of vehicles accumulate every week and retraining on it
- Vehicle OS and SDV software engineer — alongside autonomous driving, the company handles the group's vehicle software platform and keeps needing OS and middleware talent
For hiring managers
Autonomous driving deep learning and data talent is sought by automaker and parts-supplier software organizations as well as Korean and global big tech companies. With the year-end target of 200 hires overlapping with real-vehicle testing, it is better to secure key leaders quickly.
For job seekers
Experience training end-to-end or VLA models on large-scale driving data and validating them in real vehicles is what this company values most right now. Experience with vehicle OS or robotics middleware is also highly useful.
An analysis written from public sources. Sources: Korea's National Pension Service public workplace data (August 2026 register), Shinailbo 2026-09-13 (media day, data collection and test vehicles, Atria AI Level 2++ in 2029), Newsis 2026-09-13 (start of VLA development, real-vehicle testing), Newspim 2026-07-23 (about 70 experienced hires, 200 hires within the year), Newspim 2026-10-02 (three external hires). Headcount is from Korea’s National Pension Service public data (Korean workplaces).
Open jobs
Banseog checks official company careers pages every day. A job is marked closed only when the official page returns a removal (404 or 410).