Staff/Principal Forward Deployed Engineer

DiDi Labs•San Jose, CA

About The Position

DiDi's autonomous driving unit was established in 2016 with the mission of developing Level 4 autonomous driving (AD) technology to make transportation safer and more efficient. In August 2019, the unit became an independent company, DiDi Autonomous Driving, dedicated to advanced AD R&D, product application, and business expansion. We believe integrating AD technology into a shared-mobility fleet will generate immense social value. By leveraging DiDi's specialized technology, operational expertise, and integrated ecosystem, we are positioned to build and operate a highly efficient, user-oriented autonomous fleet. At DiDi Autonomous Driving, we firmly believe that the future of mobility goes beyond simply "utilizing AI"—it will be fundamentally reimagined and entirely driven by an AI-Native architecture . We are seeking a visionary, highly technical, and mission-driven Staff / Principal Forward Deployed Engineer (FDE) to act as the ultimate catalyst for our company-wide AI transformation. In this strategic, high-impact role, you will combine cutting-edge Large Language Model (LLM) expertise, robust systems architecture design, and a proven track record of enterprise-level AI scaling. You will embed deeply with our core engineering teams to evolve our traditional R&D organization into a truly AI-Native powerhouse.

Requirements

A proven technical leader who can design complex, system-level architectures while maintaining a fierce passion for writing core code, debugging deep system issues, and optimizing low-level execution paths.
Proficiency in core languages such as C++, Python, Java, JavaScript, etc.
Demonstrated ability to build technical authority, align priorities, and drive diverse engineering teams (Algorithms, Infrastructure, Hardware) toward adopting an AI-first engineering paradigm without relying on formal administrative authority.
Proven experience leading or heavily contributing to a large-scale corporate "AI-native transformation," or a track record of building enterprise-grade AI/ML platforms from 0 to 1.
Thorough hands-on deployment, tuning, and optimization experience with mainstream AI infrastructure tools and frameworks, including but not limited to PyTorch, Ray, vLLM, Triton Inference Server, Kubernetes, DeepSpeed, and Megatron-LM.
Years of deep, practical experience in distributed LLM training/inference optimization and large-scale compute cluster infrastructure & operations (I&O).

Nice To Haves

Familiarity with autonomous driving algorithms (Perception, Planning, Control, Simulation), robotics, physics-based simulation engines, or ultra-large-scale ML training/serving clusters is highly preferred.
8-10+ years of professional engineering depth in systems software, core cloud infrastructure, or production-grade machine learning platforms.
Hands-on experience building custom AI Copilot applications, autonomous Multi-Agent Frameworks, or high-tier developer productivity platforms.
Proven success steering core project delivery amidst complex business logic, fast-paced/high-pressure environments, or mission-critical systems.
An active contributor to the broader tech community (e.g., open-source maintainer/owner, author of high-quality technical blogs/papers, or speaker at premier industry AI/ML conferences).

Responsibilities

Spearhead the evaluation, selection, and deep integration of frontier LLM ecosystems (e.g., Llama, Hugging Face) and commercial AI platforms.
Own the architectural design of our unified, distributed AI platform spanning complex data processing, model training, inference pipelines, and evaluation frameworks.
Embed directly with core autonomous driving teams (Perception, Prediction, Planning & Control, and Simulation) via the FDE model.
Pinpoint engineering bottlenecks, eliminate friction, and translate complex AI capabilities into production-ready internal ecosystems (e.g., AI DevOps, AI Copilots).
Design and implement highly resilient, scalable automation pipelines for LLM deployment, monitoring, and continuous feedback loops.
Optimize GPU cluster utilization, minimize inference latency, and maximize throughput across large-scale production environments.
Keep a strong pulse on breakthrough trends in AGI and systems engineering.
Act as a "super-connector" between external technological innovations and internal systems, ensuring our AI infrastructure maintains a 1-3 year competitive edge.