Physical AI Researcher

Apply Send to a friend

---

(Note: Japanese candidates are very welcome! This job description is in English because English is the main language of research today, and also of our department, even though Japanese is the main language at Cypremos overall.)

We are a fast-growing startup founded in November 2024 as a wholly owned subsidiary of Astemo, a Japanese Tier 1 automotive supplier. We specialize in software and cloud services, and our goal is to support global automotive manufacturers’ transition to Software-Defined Vehicles (SDV) and to enable new mobility services.

Our core offerings are digital development environments that use cloud technologies such as digital twins and artificial intelligence. As cyber technologies like IoV, SDV, cloud and AI evolve, we aim to redefine how OEMs and suppliers work and how end users experience mobility, ultimately offering a kind of freedom no one has imagined before.

We are a small research team within this startup, with two missions: (A) understanding and developing new technologies that support and extend future business, and (B) giving our development teams technological guidance, based on a fundamental understanding of each topic and of the technologies available now and expected in the future. Conventional thinking says that more complex technology needs larger research teams. We challenge that: we believe a small team of highly autonomous researchers, close to industry needs but allowed high flexibility, amplified by AI and open collaboration, can explore more freely, move faster, challenge conventional beliefs while staying true to ethical convictions and see further than a large organization.

In this position, you will help us start our Physical AI efforts from the ground up. Physical AI means AI that perceives, reasons and acts in the physical world: vehicles, robots, and systems that may or may not share space with people. You will shape our strategy in this field, build our first research environment, and find the important problems that others have overlooked. In particular, you will look at how physical AI systems can act safely, appropriately among each other and socially in a world full of humans.

If you want to explore ideas that expand human potential, question them boldly, refine or overturn them when necessary, and keep evolving them; if you want to do this in a small team, on your own initiative, with external researchers you choose yourself; and if you also share our belief in the ethical application of technology, we would love to have you on board.

Required Skills & Experience

At least three years of experience in research or development related to robotics, autonomous driving, control, or machine learning for physical systems, or equivalent experience at the level of graduate research.
Hands-on experience with at least one of the following: simulation (e.g. Isaac Sim, MuJoCo, CARLA, Gazebo), robot software (e.g. ROS 2), reinforcement learning or imitation learning, or perception and sensor fusion.
Strong thinking and writing skills: the ability to structure complex ideas and communicate them clearly in writing.
Ability to do research, gather information, and discuss both ideal concepts and real-world constraints in English. Japanese is a plus.

Preferred Skills & Experience

Experience with foundation models, vision-language-action models or world models applied to robots or vehicles.
Experience bringing a system from simulation to real hardware, and dealing with the gap between the two.
Experience starting a new research area, lab or project from scratch, for example in a startup.
Interest in human–robot interaction, social navigation, or how physical AI systems behave appropriately around people.
Experience building prototypes with sensors, actuators, embedded systems or small robots.
Experience or networks in the automotive, robotics or mobility industries.
Experience researching or writing patents and academic papers.

Ideal Candidate Profile

Enjoys building things from zero and is comfortable making decisions when there is a lot of uncertainty.
Balances curiosity about new trends with clear thinking about where we can make a real difference.

Survey the fast-moving Physical AI field, for example robot foundation models, vision-language-action models, world models and sim-to-real transfer, and identify which important problems are still unsolved.
Propose a research strategy and roadmap for our Physical AI efforts, and decide where and when we should enter.
Set up our first research environment, including simulation, data pipelines and small hardware platforms, starting with a lean budget and scaling as results come in.
Design, carry out and evaluate experiments and prototypes, from simulation to real-world tests.
Connect Physical AI research with our existing strengths: digital twins, cloud-based development environments, autonomous driving, and modeling human and social behavior.
Find and build relationships with external partners, such as universities, research institutes, startups and hardware vendors.
Publish research results as academic papers and patents.
Transfer the results and knowledge to development teams, and give executives strategic advice on Physical AI trends.

We aim to measure our output by meaningful publications and patents, so what matters most to us is your ability to think deeply, question assumptions, and communicate ideas clearly in writing. A degree in a related field (for example computer science, AI, human factors, cognitive science, psychology or robotics), published papers or filed patents are welcome. But if you can show strong thinking and writing skills in other ways, such as essays, blog posts, technical reports, open-source work or independent projects, we would be just as happy to hear from you.

We want our research to be not only profitable but also meaningful, thoughtful, well executed and beneficial to society.

With your CV, please send us a short description of your past research or projects, along with samples of your work. Examples include publications, Master’s or doctoral theses, patents, technical reports, essays, blog posts or open-source projects. You can attach PDFs or share links. We will use them only to evaluate your application. They help us speed up screening.

As fellow researchers, we would also love to know what motivates your work and what kind of world you dream of and are working to make real.

---

コアタイム無しのフルフレックス、もしくは裁量労働制適用
Full flex-time system with no core hours, or a discretionary work system

社会保険完備、通勤交通費実費支給、給与改定年1回、試用期間3か月（条件変更なし）
Social insurance provided, commuting expenses reimbursed at actual cost, annual salary review, 3-month probationary period (no change in terms)

土日休日
年間休日121日
カレンダーは会社カレンダーに基づく
有給休暇有り

Saturdays and Sundays are holidays
Annual holidays total 121 days
The calendar follows the company calendar
Paid leave is granted

Hybrid (remote + office)

Country / Region

Japan

Prefecture (State)

Tokyo

Location (Address)

渋谷区

桜丘町1番地1号

渋谷サクラステージSHIBUYAタワー37F

---

Apply Send to a friend

Astemo Cypremos is a software and cloud services company in the SDV domain.
Within the SDV platform, it drives the transformation of digital engineering in the Out-Car domain beyond In-Car (in-vehicle systems), and aims to realize flexible and wide-ranging systems as a supplier that delivers the value required in the SDV era through next-generation mobility solutions leveraging cloud and communication technologies.

---