Jobs at DiDi

Lead the architecture and evolution of infrastructure powering autonomous driving software, including low-latency IPC middleware, runtime engines, compiler and build toolchains, simulation platforms, crash diagnostics, and developer productivity systems. Drive performance profiling, memory optimization, fault isolation, and cross-team engineering standards while mentoring engineers and guiding technical architecture across autonomy and simulation teams.

Lead company-wide AI transformation by designing unified distributed AI platforms, integrating frontier LLM ecosystems, and embedding with autonomous driving engineering teams. Build scalable LLMOps and MLOps pipelines, optimize GPU infrastructure and inference performance, develop AI-native engineering tools, and establish technical roadmaps. The role combines hands-on systems engineering, enterprise AI platform development, cross-functional influence, and strategic leadership across perception, prediction, planning, control, and simulation teams.

Conduct research on scene understanding and planning for autonomous vehicles. Design, train, evaluate, and debug deep learning models using large-scale driving data and sensor inputs. Develop metrics, perform failure analysis, survey research, and integrate validated prototypes into production planning systems while collaborating with perception, prediction, planning, mapping, and engineering teams.

Develop ML-powered motion planning systems for Level 4 autonomous vehicles. Responsibilities include building learned planners and end-to-end models, applying imitation and reinforcement learning, integrating ML with rule-based and optimization-based planning, training and evaluating models using real-world data and simulation, and deploying production systems under safety, latency, and onboard compute constraints. The role also involves scenario mining, evaluation infrastructure, closed-loop simulation, and collaboration across autonomous driving teams.

Lead safety strategy and deployment readiness for fully autonomous driving systems. Define safety requirements, system architectures, release policies, degradation and fallback criteria, V&V plans, operational safety workflows, emergency protocols, and measurable safety metrics. Analyze incidents during unmanned operations and drive improvements across system design, testing, and procedures. Collaborate with Software, Product, Operations, and Safety teams to support autonomous deployment, city expansion, and operational scaling.

Develop core decision-making and motion planning systems for Level 4 autonomous vehicles. Design behavioral planning, path and velocity generation, trajectory ranking, world modeling, and optimization algorithms. Analyze and debug performance across driving scenarios while collaborating with Prediction, Perception, and Control teams. Implement real-time algorithms in C++ to enable safe, comfortable, and intelligent vehicle navigation.

Develop core decision-making and motion planning systems for Level 4 autonomous vehicles. Responsibilities include behavioral planning, path and velocity generation, trajectory optimization, world modeling, trajectory cost design, testing, debugging, and root-cause analysis. The engineer will collaborate with Prediction, Perception, and Control teams to deliver safe, comfortable, and intelligent vehicle motion in complex environments.

Lead deployment, inference optimization, resource scheduling, profiling, and stability initiatives for AI models across autonomous vehicles and cloud infrastructure. Design low-latency inference pipelines, optimize GPU execution and memory usage, support LLM serving and model validation, diagnose software-hardware integration issues, and collaborate with perception, prediction, cloud infrastructure, and safety teams.

Develop motion and behavioral planning algorithms for Level 4 autonomous vehicles. Responsibilities include lane-change and merge decision-making, trajectory optimization, geometry and velocity planning, world modeling, cost-function design, simulation and real-world testing, debugging, and collaboration with perception, prediction, and control teams. The role requires PhD-level research experience in robotics or related fields, strong C++ skills, and publications in leading conferences.