Job DescriptionTranslated from Japanese
▼Job Description
Our company's vision is to "Change the world with innovation," and we support the social implementation of AI. Founded in 2012, we have been at the forefront of AI, and by leveraging our accumulated knowledge and actively adopting new technologies, we continue to maintain a high growth rate even after our IPO.
We provide an integrated platform and solutions from data analysis to AI model development and operation, realizing on-site problem-solving with AI. Centered around "ABEJA Platform," a suite of software for continuous and stable operation, we are developing a digital platform business that promotes DX for our clients. We have supported the DX of over 300 companies across various industries, including manufacturing, logistics, retail, finance, and infrastructure.
We are looking for colleagues who can help us achieve further business growth and solve our clients' management and business challenges!
▍Background of Recruitment
ABEJA is advancing the social implementation of Physical AI technologies, including Vision-Language-Action (VLA), as actual robot systems at client sites.
While hardware engineers, middleware engineers, and data scientists each demonstrate their expertise, the importance of the software layer that spans these disciplines and makes the entire system cohesive is increasing.
We are seeking engineers who can lead software development that shapes the value of Physical AI, including data collection infrastructure from robots, service layers that collaborate with learning and inference, and applications integrated into actual hardware.
▍Mission
The mission of this position is to design and implement the entire software layer for Physical AI systems, realizing a state where robots, data, and AI are consistently connected.
You will technically lead projects by understanding the constraints of hardware and middleware, and by building the software infrastructure that supports everything from data collection to learning, inference, and on-site operation.
▍Job Responsibilities
- Designing the software architecture that connects hardware, middleware, AI models, and applications, with an overview of the entire robot system.
- Designing and implementing data collection infrastructure and pipelines for stable collection and management of sensor data and logs acquired from robots.
- Designing backend services and APIs that collaborate with learning and inference systems, and connecting them with actual robots and higher-level applications.
- Developing applications and service layers integrated into robots, and implementing and improving functions with on-site operation in mind.
- Designing infrastructure configurations, including cloud and on-premises, and establishing the foundation to support development, verification, and operation.
- Abstracting developed content for each client and developing common products.
- Flexibly reviewing and improving software design and implementation based on requirements and constraints identified during on-site verification and client implementation.
- Collaborating with hardware engineers, middleware engineers, data scientists, and project managers to lead technical decision-making and design direction.
▼Required Skills/Experience
- Practical experience leading software development from design to implementation using at least one language such as Python, Java, Go, or C++.
- Experience in developing service layers, including web services, backend systems, and API design.
- Experience in development and operation in a Linux environment, with practical experience in process management, debugging, and log analysis.
- Experience in organizing requirements and making design decisions while collaborating with multiple technical domains and stakeholders.
- Self-driven ability to think independently, form hypotheses, and iterate through implementation and verification, even in situations with unorganized requirements or high uncertainty.
▼Welcome Skills/Experience
- Experience in software development that interacts with physical systems, such as robots, IoT, or edge devices.
- Experience in designing and implementing data collection infrastructure, streaming processing, and distributed systems.
- Experience in integrating learning and inference pipelines into products in collaboration with machine learning and data science teams.
- Experience in designing and building development and operation infrastructure, such as cloud infrastructure, containers, and CI/CD.
- Experience leading a team in technology selection and architecture design.
▼Ideal Candidate Profile
- Individuals who find value in considering how to make a Physical AI system cohesive, rather than focusing solely on individual technical components.
- Individuals who find satisfaction in connecting different specialized domains such as hardware, middleware, data, and AI, and acting as a technical translator.
- Individuals who can move forward by iterating between hypotheses and implementation, even in situations with incomplete prerequisites.
- Individuals who enjoy the challenge of supporting social implementation with the power of software in the nascent field of Physical AI.
▼Job Highlights
- Experience in taking a central role in software design that integrates robots, data, and AI in the new technological domain of Physical AI.
- An environment where you can be involved in end-to-end system design, including data collection, learning, inference, and on-site operation, not just single application development.
- Practical experience in designing and implementing software that can withstand the real world, considering the constraints of hardware and middleware.
- The opportunity to work with discretion as a lead engineer in a phase where technical decisions and designs directly impact the success of the entire project.
- Experience supporting the integration of Physical AI into actual client operations from a software perspective.
▼Development Environment
▍Development Environment
- Development Languages: Python, C++
- Web Backend: FastAPI
- Machine Learning: NumPy, SciPy, scikit-learn, OpenCV, Pinocchio, PyTorch, Isaac Lab
- Development Tools: CMake, ament, bloom, colcon, pixi, uv
- Frameworks: LeRobot
- Version Control: Git, rosdep, vcs
- Middleware: ROS 2, Zenoh Router
- Containers/Virtual Machines: Docker, VirtualBox
- Simulation: RViz2, MuJoCo, IsaacSim
- Databases: SQLite 3+, MySQL 8+, mcap
- CI/CD: GitHub Actions
- IaC: Terraform
- Logging/Telemetry: OpenTelemetry, Datadog
- Robots Used: AGIBot G2, OpenArm, Amazing Hand, SO-ARM, others
▼Employment Type
Full-time employee (No fixed term)
▼Probationary Period
3 months
▼Salary
Annual salary: ¥6,000,000 - ¥12,000,000
▼Minimum Annual Salary (in 10,000 JPY)
600
▼Maximum Annual Salary (in 10,000 JPY)
1200
▼Work Location
1-1-14 Mita, Minato-ku, Tokyo 108-0073 Bizflex Azabu-Juban 2F
*Expected to come to the office about 3 days a week.
【Scope of Change】
Location determined by the company (including locations for remote work)
▼Working Hours
＜Flextime System＞
10:00 - 19:00 (Standard working hours: 8 hours per day)
*Core time: 10:00 - 15:00
*Flexible time: 07:00 - 10:00 / 15:00 - 19:00
▼Break Time
60 minutes
▼Overtime Work
Yes
▼Holidays, Paid Leave, Special Leave
Complete two-day weekend (Saturday, Sunday), National Holidays
Annual paid leave (10 days in the first year)
Refresh leave (up to 3 days in the first year)
Others (Bereavement, Golden Week, Year-end and New Year holidays, Childcare leave, Family care leave, etc.)
▼Benefits
- Child Allowance: ¥10,000 per month per child up to 20 years old, paid after the probationary period.
- Health Check-ups: Conducted once a year for those enrolled in health insurance.
- Growth Support System: Purchase of necessary content (books, e-books, online courses, etc.) for achieving one's mission is possible with company expenses without prior approval.
- Marriage/Childbirth Celebration Money: ¥30,000 paid as a congratulatory gift upon employee's marriage or birth of a child.
*Several other systems are also in place.
▼Allowances
- Midnight and Holiday Work Allowance
- Additional Premium Allowance
- Overtime Work Allowance
*Transportation expenses will be reimbursed based on actual costs.
▼Insurance Enrollment
Various insurances: Health Insurance, Employees' Pension Insurance, Employment Insurance, Workers' Accident Compensation Insurance
Health Insurance: Enrolled in the Kanto IT Software Health Insurance Association (can use affiliated facilities, etc.)
▼Measures to Prevent Passive Smoking at the Workplace
No smoking on premises
▼Health
Health Check-ups: Conducted once a year for those enrolled in health insurance.
▼Interview/Selection Process
- Approximately 3-4 interviews are expected.
- The selection period is approximately 3 weeks.
- Interviews are generally conducted online. (We are flexible if you prefer offline interviews.)
▼Other
▍ABEJA actively shares technical information!
note: https://note.com/abeja
Tech Blog: https://note.com/abeja
Tech LT: https://abeja-innovation-meetup.connpass.com/