4 days ago
Shanghai, ChinaStaff+
Responsibilities
- Develop multi-object tracking with data association, Bayesian state estimation, track lifecycle management, multi-sensor association, and learning-based tracking methods.
- Build generic and open-set world understanding using class-agnostic obstacle detection, occupancy representations, clustering, and learned generic-object branches.
- Develop motion prediction for pedestrians, cyclists, vehicles, and other agents, including multimodal forecasting, interaction-aware prediction, occupancy flow, and uncertainty estimation.
- Evaluate and productionize temporal perception-prediction systems, including joint detection-tracking-forecasting, streaming BEV representations, trajectory queries, and learned world models.
- Explore and adapt VLM, VLA, and embodied foundation models for open-vocabulary perception, semantic reasoning, auto-labeling, and task-conditioned robot intelligence.
- Integrate camera, LiDAR, radar, and IMU data while maintaining geometric, temporal, identity, prediction, and semantic consistency.
- Define tracking, prediction, occupancy, safety, and embedded-latency quality gates and perform fleet-log root-cause analysis.
- Partner with Detection, Prediction, Planning, Data & Infra, and Integration teams to move systems from research into reproducible training, replay, simulation, and robot testing.
Requirements
- At least 3 years of hands-on experience shipping or operating perception or prediction systems for robotics, autonomous vehicles, or ADAS, with strong autonomous-driving experience.
- Rich experience with classical/geometric and learning-based approaches and judgment regarding modular, end-to-end, and hybrid designs.
- Proven depth in multi-object tracking, including data association, Bayesian filtering, track lifecycle, temporal and multi-sensor association, ID stability, and low-latency evaluation.
- Hands-on experience with generic or open-set perception, occupancy representations, LiDAR clustering, anomaly or unknown-object handling, or learned-detection fusion.
- Hands-on depth in a modern temporal area such as motion forecasting, occupancy flow, streaming or temporal BEV, learned world models, or joint detection-tracking-prediction.
- Practical understanding of geometry, multi-sensor calibration, time synchronization, reproducible experimentation, simulation, replay, and log-driven root-cause analysis.
- Proficiency in C++ and Python for production-quality code, training and evaluation tools, and real-time embedded performance analysis.
- Preferred experience with motion forecasting, VLM/VLA or embodied foundation models, foundation-model adaptation, 3D or BEV detection, sensor fusion, robotics libraries, edge inference, and model lifecycle operations.
- Demonstrated ability to use AI tools and systems to accelerate engineering and research while maintaining quality, security, and reproducibility.
Benefits
- Term Life Insurance and comprehensive Medical Insurance.
- GrabFlex benefits package flexibility.
- Parental leave, Birthday leave, and Love-all-Serve-all volunteering leave.
- Confidential Grabber Assistance Programme.
- Onsite work at a Grab office in Singapore or China.
Categories
ML EngineeringRobotics
About Grab
Grab builds a Southeast Asia superapp for mobility, deliveries, and financial services, serving consumers, drivers, merchants, and enterprises. It monetizes via ride and delivery commissions, merchant services (such as ads and subscriptions), and payments and digital banking products. Founded in 2012 and headquartered in Singapore, Grab operates across multiple Southeast Asian markets and is publicly listed on NASDAQ.