上车 AV 软件工程师
Onboard AV Software Engineer
关于Humble Robotics
在Humble Robotics工作意味着参与几十年来地面运输领域最大的变革。我们正在打造一款自动驾驶、零排放的货运车，通过突破性的基于视觉的AI大幅降低货运成本，专为当今全球物流网络设计。
我们是一支快速前进、紧密协作的团队，成员包括自动驾驶行业资深人士和富有创新思维的人才。我们认为文化无法被设计出来——但当它自然形成时，这将是一次终生难忘的冒险。
进步从未如此贴近现实。
职位概述
我们正在寻找一名软件工程师，负责在卡车的车载计算设备上优化和部署机器学习模型，并负责整个自动驾驶系统的性能。你将从我们的机器学习团队获取模型，并使其在嵌入式GPU上快速、高效且可靠地运行——使用TensorRT、自定义CUDA内核和底层系统工程。除了推理之外，你还将对整个车载软件流程进行分析和优化，以满足严格的实时性要求。这是一个难得的机会，将机器学习与嵌入式系统结合，用于生产级自动驾驶货运，伴随着小团队解决重大问题所带来的自由与责任。
主要职责
- 使用TensorRT和自定义CUDA内核，对车载推理的神经网络模型进行优化和部署
- 分析并减少自动驾驶系统中的端到端延迟——从传感器数据采集到控制
- 构建和维护车载C++和Rust软件基础设施，包括实时数据管道、进程间通信和硬件抽象层
- 实现模型量化、剪枝和其他优化技术，以最大化嵌入式GPU平台的吞吐量
- 与机器学习工程师合作，确保模型设计适合高效部署，并与车辆系统工程师合作，满足实时安全约束
最低要求
- 计算机科学、电气工程、机器人学或相关领域的学士、硕士或博士学位——或同等行业经验
- 在高性能关键系统中熟练使用C++和/或Rust
- 具有GPU加速计算的实际经验——CUDA、TensorRT或类似的推理优化工具链
- 熟悉机器学习模型架构（如transformers、CNNs），并能够分析计算成本和内存占用
- 有资格在美国工作
薪酬
该职位提供基本薪资+福利+股权补偿。
查看英文原文
About Humble Robotics
Working at Humble Robotics means taking on the biggest change in ground transportation in decades. We’re building an autonomous, zero-emissions hauler that dramatically lowers the cost of freight with groundbreaking vision-based AI, designed for today’s global logistics network.
We’re a fast-moving, close-knit team of AV industry veterans and innovative thinkers. We don’t believe culture can be engineered – but when it falls into place, it’s a once-in-a-lifetime adventure.
Progress has never felt so present.
Position Overview
We're looking for a software engineer to optimize and deploy ML models on our trucks' onboard compute, and to own performance across the full autonomous driving stack. You'll take models from our ML team and make them run fast, efficiently, and reliably on embedded GPUs—using TensorRT, custom CUDA kernels, and low-level systems engineering. Beyond inference, you'll profile and optimize the entire onboard software pipeline to meet hard real-time deadlines. This is a rare chance to bridge ML and embedded systems for production autonomous freight, with the freedom and responsibility that comes with a small team tackling a massive problem.
Key Responsibilities
- Optimize and deploy neural network models for onboard inference using TensorRT and custom CUDA kernels
- Profile and reduce end-to-end latency across the autonomous driving stack—from sensor ingestion to control
- Build and maintain the onboard C++ and Rust software infrastructure, including real-time data pipelines, inter-process communication, and hardware abstraction layers
- Implement model quantization, pruning, and other optimization techniques to maximize throughput on embedded GPU platforms
- Collaborate with ML engineers to ensure models are designed for efficient deployment, and with vehicle systems engineers to meet real-time safety constraints
Minimum Qualifications
- BS, MS, or PhD in Computer Science, Electrical Engineering, Robotics, or a related field—or equivalent industry experience
- Strong proficiency in C++ and/or Rust for performance-critical systems
- Hands-on experience with GPU-accelerated computing—CUDA, TensorRT, or similar inference optimization toolchains
- Familiarity with ML model architectures (transformers, CNNs) and the ability to reason about computational cost and memory footprint
- Eligible to work in the United States
- Experience with onboard software for autonomous vehicles, robotics, or IoT/edge devices
- Deep knowledge of CUDA, TensorRT, model quantization, and kernel-level optimization
- Experience with Bazel or similar build systems for complex codebases
- Familiarity with real-time robotic systems
- Experience profiling and optimizing full-system performance (CPU, GPU, memory, I/O) on embedded platforms
- Comfort operating as an early team member—high ownership, low ego, fast iteration
Compensation
This role is eligible for base salary \+ benefits \+ equity compensation. Salary ranges are determined by role, level, and location. Within the range, individual pay is determined by additional factors, including qualifications, skills, experience, and location.
Additional Information
As part of the interview process, we may use Artificial Intelligence (AI) tools to compare your qualifications and experience to the job description. A human reviews all AI output and makes a final hiring decision. Humble Robotics does not rely on the output to make any employment decisions. Some applicants may have a legal right to opt-out of the use of AI as part of our interview process. Contact **legal@humblerobotics.ai** to exercise this right or if you have further questions on the use of AI tools in our hiring process.
Humble Robotics is an equal opportunity employer. All qualified applicants will receive consideration for employment without regard to race, national origin, gender, age, religion, disability, sexual orientation, veteran status, marital status or any other characteristics protected by law. Humble Robotics will consider qualified applicants with arrest and conviction records in a manner consistent with local ordinances.