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.
This role develops learned driving policies that connect multimodal observations, vehicle state, route context and, where applicable, language instructions to actions or trajectories. You will work with research, planning, controls and integration colleagues to turn model improvements into reproducible evidence and integration-ready artifacts. Own a well-defined model development workstream, from data preparation to evaluation and a reproducible handoff, for end-to-end driving or vision-language-action (VLA) policies.
JOB DESCRIPTION
- Build synchronized training examples from observations, vehicle state and demonstrations; prevent future-information leakage and keep training and evaluation groups separate.
- Train or adapt a policy model with supervised or imitation learning; compare against a baseline and record configurations, data versions and controlled experiments.
- Evaluate predictions and closed-loop behavior on assigned scenarios; investigate collisions, constraint violations, route progress, comfort and distribution shifts.
- Implement and test agreed input/output contracts with planning, controls and integration colleagues, including coordinates, timestamps and rejection of invalid or stale outputs and instructions.
- Package model weights, preprocessing, tokenizer where used, and regression results; measure latency and observation age, and help resolve integration failures.
REQUIREMENTS
- A bachelor’s degree in computer science, AI, robotics, electrical engineering or a related field, or equivalent practical experience demonstrated through relevant work.
- Hands-on Python and a deep learning framework such as PyTorch or JAX; ability to train, debug, save and evaluate models reproducibly.
- A completed project with a learned robotics policy, temporal vision model or VLM; ability to explain your contribution, data choices, baseline and failure cases.
- Working knowledge of optimization, overfitting, sequence modeling and causal attention; ability to check temporal alignment and evaluation leakage.
- Understanding of coordinate frames, motion representations and the difference between policy trajectories and actuator commands; ability to implement and test model interfaces.
- Ability to use Git, document experiments, read technical English and communicate model limitations to collaborators.
Preferred skills
- Experience with E2E or VLA policies, imitation learning, or closed-loop robotics simulation.
- Experience with parameter-efficient adaptation such as LoRA, or other model adaptation methods supported by comparative results.
- Experience with multi-camera or temporal data, model distillation, quantization, or inference profiling.
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)