3 months ago
Mountain View, CA, USASenior
Base Salary
$190k - $260k/yr
Responsibilities
- Design high-throughput data loading and streaming systems for multimodal camera, LiDAR, and radar data, including dataset formats, sharding, and prefetching.
- Build and optimize distributed training infrastructure across multi-node GPU clusters using data, tensor, pipeline, FSDP, and ZeRO parallelism.
- Maximize accelerator utilization through mixed-precision training, fused kernels, memory optimization, and communication/computation overlap.
- Profile end-to-end training pipelines and eliminate storage, network, CPU preprocessing, and GPU-compute bottlenecks.
- Develop scalable pipelines that convert petabytes of raw driving logs into streamable training datasets.
- Partner with ML teams to scale new architectures from prototypes to reliable full-cluster training runs.
Requirements
- BS, MS, or PhD in Computer Science or a related field.
- At least 2–3 years of industry experience in ML systems or infrastructure.
- Hands-on experience with PyTorch DDP/FSDP, DeepSpeed, Megatron, NCCL, and distributed-training techniques.
- Experience building high-performance large-scale training data pipelines using streaming formats such as WebDataset or MosaicML Streaming/MDS, with sharding and storage/network-aware loading.
- Deep understanding of GPU performance, including mixed precision, memory hierarchy, kernel fusion, Nsight, PyTorch Profiler, NVLink, and InfiniBand.
- Strong Python skills and proficiency with PyTorch internals.
- C++/CUDA/Triton systems-level experience is a plus.
Benefits
- Competitive compensation package including equity and annual bonuses.
- Medical, dental, and vision plans through Kaiser Permanente, Cigna, and MetLife, including infertility benefits.
- Legal services, identity and fraud protection, hospital indemnity, accident, critical illness, disability, and life insurance.
- Flexible PTO, 10 paid holidays, and generous parental leave.
- Mountain View, California office with dog-friendly facilities, catered lunch, stocked kitchen, and free EV charging.
- Headspace, Calm, One Medical, Gympass, Spring Health, and Rula wellbeing benefits.
- Fidelity 401(k), commuter benefits, FSA, dependent care FSA, HSA, referral bonuses, and patent bonuses.
- Visa sponsorship may be available for eligible candidates, subject to stated restrictions.
Categories
About Kodiak
Kodiak builds an AI-powered autonomous driving stack for long-haul trucking and government programs. The company hauls freight daily across the southern United States using its technology and is leveraging the same software for U.S. Department of Defense autonomy projects. Founded in 2018 and headquartered in Mountain View, California, Kodiak announced in 2024 that it had delivered a driverless semi-truck to a customer.