INTRODUCTION
Phenikaa-X is a Phenikaa Group company developing autonomous vehicles, robotics and automation technology in Vietnam. Our portfolio includes autonomous vehicles, autonomous mobile robots and drones, supported by technologies such as artificial intelligence, mapping and simulation. We aim to bring autonomous technology into transportation, manufacturing and service operations, with a focus on safety, reliability and operational efficiency.
Our research team develops methods for perception, prediction and planning in autonomous driving, with priority directions in world models, foundation models and reinforcement learning. Researchers focus on a relevant domain and work with ML, simulation and integration engineers to test its contribution to the system.
You will lead a substantial research problem in perception, prediction or planning. In your chosen domain, you will select modeling and post-training methods, including world models, foundation models or RL where appropriate, and turn experiments into validated system improvements.
JOB DESCRIPTION
- Design a research program around a concrete perception, prediction or planning limitation; choose representations and architectures, including world or foundation models where appropriate, with measurable system targets.
- Lead studies of learning and post-training strategies appropriate to your domain, such as RL, distillation or fine-tuning; test training objectives and data mixtures for downstream performance.
- Design distributed experiments with ML engineers, select DDP, FSDP or sharding where appropriate, and resolve training instability, memory bottlenecks and reproducibility failures.
- Define offline and closed-loop comparisons, scenario coverage and regression checks; investigate compounding rollout errors and simulation-to-system gaps before recommending integration.
- Mentor researchers, review critical code and experiments, and hand over validated methods with engineers; lead technical writing and publications allowed by company policy.
REQUIREMENTS
- A master’s or PhD in ML, computer science, robotics or a related field, or equivalent research experience. Bachelor’s holders with a strong research portfolio are eligible; bring at least 3 years of relevant research, including graduate research.
- A record of leading research in an autonomy domain, world models, foundation models or RL from problem definition to a validated result; show original contributions and decisions through papers, technical writing or substantial open-source work.
- Deep understanding of one relevant model family and its training dynamics, with practical experience in post-training, policy optimization or distillation.
- Strong Python and PyTorch or JAX skills; experience designing distributed training with DDP, FSDP or comparable sharding and diagnosing memory, throughput and convergence problems.
- Experience designing ablations and closed-loop or sequential-decision evaluations, mentoring researchers and explaining method choices and limitations to engineering teams in English.
Preferred skills
- Primary contributions to research published at leading ML or robotics venues, or sustained ownership of a widely used research codebase.
- Experience building on-policy RL or preference-learning pipelines, including rollout generation, reward validation and policy evaluation.
- Experience adapting diffusion, flow-matching or autoregressive world models to driving data and reducing inference cost through distillation.
- Experience with multi-node training, tensor-parallel inference or transferring research into a robot or vehicle software stack.
BENEFITS
- Salary: negotiation based on candidate’s capacity.
- Lunch allowance, holiday bonuses, 13th-month salary.
- Birthday, sick leave, maternity, and welfare policies according to Vietnam Labor Code.
- Overtime pay in accordance with legal regulations (150–300% rate depending on time).
- Full social insurance contribution; private health insurance for employees and managers.
- Dynamic, friendly working environment with passionate and talented engineers.
- Equipped with modern equipment such as a high configuration PC.
- Accessed high-configured supercomputers for training AI models.
CONTACT
Recruitment Department – Phenikaa-X JSC
- Email: tuyendung@phenikaa-x.com
- Contact: 0971131001 (Ms. Thuy Duong)