# Software Engineer - GPU Kernels

**Company:** [Baseten](https://hotfix.jobs/companies/baseten)
**Location:** San Francisco, CA, New York, NY
**Role:** Backend Engineering
**Salary:** $180k – $360k/yr
**Skills:** CUDA, C++, Ptx, Nsight Systems, Nsight Compute, Torch Profiler, Tensor Cores, Flash Attention, Cutlass, Triton
**Posted:** 2025-07-17

> Develops and optimizes high-performance GPU kernels for AI/ML operations like matrix multiplications and attention mechanisms using CUDA and C++. Requires deep GPU architecture knowledge, performance profiling, and low-level optimization expertise.

## Job Description

## Responsibilities

### Core Engineering Responsibilities
- Design and implement high-performance GPU kernels for key ML operations, including matrix multiplications, attention mechanisms, and mixture-of-experts routing
- Write and optimize code using **CUDA**, **PTX assembly**, and architecture-specific techniques
- Apply advanced performance optimization methods such as memory coalescing, warp-level programming, tensor core acceleration, and compute/memory overlap

### Performance & Innovation
- Implement cutting-edge features like quantization (**FP8/FP4**), sparsity, and compute/communication overlap
- Identify and resolve performance bottlenecks using tools like **Nsight Systems**, **Nsight Compute**, and **Torch Profiler**
- Collaborate with research teams to productionize theoretical advancements

### Impact & Collaboration
- Contribute to internal and open-source GPU libraries
- Present technical contributions at industry conferences (e.g., **NVIDIA GTC**, **AWS re:Invent**)

## Requirements
- Strong understanding of GPU architecture and programming paradigms:
  - Memory hierarchy (global, shared, registers, **L1/L2 cache**)
  - Thread/block/grid organization
  - Synchronization techniques and race condition mitigation
- Proficient in **C++** and GPU performance profiling tools
- Knowledge of:
  - **CUDA C++ API**
  - Memory access patterns and bandwidth optimization
  - Numerical precision and quantization strategies
  - Modern GPU features (e.g., tensor cores, async operations)

## Nice to Have
- Experience with Transformer models and attention optimization (e.g., **Flash Attention**)
- Familiarity with GPU kernel libraries: **Cutlass**, **Triton**, **Thrust**, **CUB**
- Background in GEMM tuning and distributed/multi-GPU compute
- Contributions to open-source GPU projects
- Research publications or conference presentations on GPU performance

## Similar roles

- [Backend Engineer](https://hotfix.jobs/jobs/8c0cdac6-4075-42b1-a874-6e856a68d2ba) - Fathom - AI Notetaker - Remote - $180k – $240k/yr
- [Software Engineer, Knowledge Systems](https://hotfix.jobs/jobs/afb75fbb-e379-420c-99a7-1eb6f7155291) - Exa - San Francisco, CA - $180k – $350k/yr
- [Software Engineer, Control Plane](https://hotfix.jobs/jobs/1f8d4743-d927-4daf-b738-f3122fe354d7) - Hightouch - Remote - $180k – $260k/yr
- [Software Engineer, Streaming Systems](https://hotfix.jobs/jobs/42426ad1-4ce5-451e-bd46-474088d85183) - Hightouch - Remote - $180k – $320k/yr
- [Software Engineer, Customer Studio Backend](https://hotfix.jobs/jobs/c1a85145-b4d5-46dd-9b35-e98f04a3d626) - Hightouch - Remote - $180k – $320k/yr

**Apply:** https://hotfix.jobs/jobs/5be2139f-3030-4b53-aee2-5c8cdc358c06
**Canonical:** https://hotfix.jobs/jobs/5be2139f-3030-4b53-aee2-5c8cdc358c06