5 hours ago
Base Salary
$170k - $282k/yr
Responsibilities
- Build agentic harnesses for autonomous driving with permissioned access to software, data, training infrastructure, and evaluation environments.
- Develop recursive self-improvement loops using failure mining, candidate generation, evolutionary search, and targeted data generation.
- Curate and validate reusable skills derived from verified fixes.
- Train and fine-tune LLMs on agent development trajectories and adapt them to the software stack, tools, and codebase.
- Develop agent-as-policy approaches that modify policy logic, curate data, adjust training recipes, and diagnose driving failures.
- Shape onboard architecture while balancing latency, compute, determinism, and safety requirements.
- Work with prediction, planning, simulation, and onboard teams to productionize research prototypes.
Requirements
- Strong machine-learning background, including deep learning, reinforcement learning, or decision-making systems.
- Hands-on experience building LLM agent systems involving tool use, multi-agent orchestration, or automated evaluation loops.
- Experience training or fine-tuning LLMs for agentic workflows, including trajectory curation, supervised fine-tuning, or reinforcement learning from execution feedback.
- Strong Python skills and working knowledge of C++.
- Rigorously conduct controlled experiments and held-out evaluations, distinguishing measured results from assumptions.
- Ability to own systems from research prototype through production.
- Preferred: experience building agentic harnesses or self-improving loops for a product or domain.
- Preferred: autonomous-driving or robotics experience, especially in planning, prediction, simulation, or sim-to-real transfer.
- Preferred: familiarity with code-as-policy, program synthesis, LLM-driven optimization, or evolutionary search.
- Preferred: publications at top machine-learning or robotics venues such as NeurIPS, ICML, ICLR, CVPR, RSS, or CoRL.
- Preferred: BS, MS, or PhD in Computer Science, Machine Learning, Robotics, or a related field.
Benefits
- Annual base salary range is $169,783–$282,264, with additional bonus, equity, and benefits.
- Compensation varies by role, level, location, skills, experience, and education or training.
Categories
About DiDi
DiDi is a mobility technology platform offering app-based ride hailing, taxi hailing, shared mobility, food delivery, and intra-city freight to consumers and businesses across Asia Pacific, Latin America, and other markets. It connects riders with drivers and couriers and provides vehicle, energy, and insurance-related services, earning transaction commissions and service fees. DiDi launched an autonomous driving unit in 2016, which became the independent DiDi Autonomous Driving in 2019 to develop Level 4 technology for its network.