SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Fractile is building specialized hardware and software systems for AI inference at scale. Founded in 2022, the company has raised $220M to tackle the fundamental bottleneck in frontier AI: inference speed. The company is developing chips and systems designed to efficiently generate outputs at thousands of tokens per second while handling the complexity of large models operating at very long contexts.
The Developer Experience team sits within Fractile's Software organisation, which is responsible for the full software stack—from ML compilers and device drivers to systems firmware, runtime, ML libraries, developer tooling, simulators, and datacenter-scale deployment solutions. The Developer Experience team specifically helps shape how customers interact with Fractile's inference accelerator by identifying, creating, and presenting information on system usage, and providing tooling to help optimize customer deployments while maintaining a shallow learning curve.
In this role, you will work across the company to understand the current state of inference hardware development and identify what tooling will resonate with engineers seeking performance improvements. You'll design tools that fit naturally into existing LLM engineer workflows with minimal disruption. This is a self-driven position requiring the ability to identify opportunities, design solutions, and deliver them in a way that integrates seamlessly with engineers' existing toolsets.
You may have previously worked as a Field Development Engineer (FDE) or at a company providing inference services using proprietary models. You'll need to develop new tools that are both functional and easy to learn and use, working with stakeholders across multiple teams.
REQUIREMENTS:
- Experience with inference frameworks such as JAX and/or PyTorch
- Experience with software development in Python, Rust, C/C++, or similar languages
- Experience monitoring and optimizing code written in CUDA, ROCm, OpenCL, or similar
- Experience monitoring and optimizing LLM deployments with 50+ billion parameters
- Experience working with stakeholders across multiple teams
