Lead Software Engineer, Perception

Job Description

About the Team:

The Robotics Technology team is a core part of Grab's long-term vision to build urban embodied AI. Our engineers take full ownership of the product lifecycle: designing and manufacturing hardware in-house, developing control and machine‑learning systems, and rigorously testing in real-world conditions and production fleet operations. We are executing an ambitious growth plan to expand our robotics fleet across cities over the coming years, and we are focused on delivering highly productive, safe and efficient robot delivery services that help address current delivery labor shortages.

Based in Singapore and China, we offer opportunities to work on the latest autonomy, deploy solutions in complex environments, and directly influence the future of last‑mile logistics. If you're excited by tangible impact, large-scale systems and cross-functional engineering, you'll find meaningful challenges and rapid career growth here.

Get to know the role

As a Lead Perception Engineer, you will lead the building and shipping of the perception stack that turns raw sensor data into reliable, real-time understanding of the world. You will be hands-on across the perception stack, contributing to shipping milestones through evaluation, debugging, integration, and iterative improvements to perception algorithms and runtime behaviour.

To accelerate early progress, you will take initial ownership of perception foundations: sensor integrity, calibration, and the tooling and validation workflows that ensure our perception inputs are geometrically and temporally correct. This work is a direct lever on perception quality and iteration speed. As these foundations stabilise, you will broaden into wider perception development, taking on larger slices of perception delivery over time.

You will report to the Senior Principal Perception & Prediction Engineer and work onsite at Grab office.

The critical tasks you will perform

• You will lead and set direction for perception capabilities across detection, tracking, and multi-sensor integration, using disciplined evaluation and failure analysis to iteratively raise quality and robustness.
• You will own perception foundations end-to-end, including sensor integrity and calibration for our multi-sensor suite (camera, LiDAR, radar, and IMU).
• You will build practical diagnostics and automated checks to catch geometric or timing issues early and prevent misdiagnosis of them as perception regressions.
• You will partner with Integration, Data, and Infra teams to ensure perception is reliable in production-like conditions, including logging and replay workflows, and performance and latency investigations.
• You will mentor engineers and raise the technical bar of the team, guiding design decisions and driving alignment across cross-functional partners.

Qualifications

Skills you need

What Essential Skills You Will Need

• At least 5 years experience shipping or operating perception systems for robotics, AV, or ADAS, with experience debugging and improving real-world performance.
• Experience working in a research setting - independently driving research end-to-end, gathering data, and presenting findings to peers including demonstrated leadership guiding a small team to a state-of-the-art result (e.g. a CVPR competition).
• Strong knowledge across machine learning, perception, prediction, planning, and simulation, including experience with end-to-end, purely image-based algorithms.
• Deep fundamentals in geometry and multi-sensor systems, including practical calibration know-how (intrinsic and extrinsic) and the ability to diagnose alignment or time-synchronisation issues that impact perception quality.
• Proficiency in C++ and Python, with the ability to write production-quality code and build lightweight tools or checks that improve iteration speed and reliability.
• Experienced navigating noisy, production-like environments, including log-driven root-cause analysis across software, configuration, and sensor behaviour.
• Engineering discipline, including reproducible workflows, clear metrics, and validation-minded execution.
• Demonstrated proficiency in leveraging AI tools and systems to drive efficiency and innovative problem-solving.

Good to have

• Familiarity with perception model lifecycle elements such as dataset curation, evaluation metrics, regression detection, deployment constraints, and inference performance optimization.
• 3+ years of experience with common robotics/perception libraries (e.g., OpenCV, PCL, Ceres) and tools for visualisation and debugging.
• Experience with sensor fusion, tracking, or 3D perception pipelines, and interest in growing deeper in these areas over time.