AI DevTools Engineer for Inference Stack

Socket.dev; Los Altos (CA); Hybrid; USD 167,000 - 242,000; Full time

Benefits offered by this job

Relocation assistance

Stock options

Premium health insurance

5% 401k matching

Job summary

Modular is seeking an engineer for the MAX Tools team to build developer tooling for its AI inference stack, enabling engineers to create, debug, and optimize models and kernels. You will work on profiling, tracing, debuggers, and dashboards, as well as agent-native tooling and integration across the stack.

The role emphasizes shipping developer tools, performance measurement, and close collaboration with kernel, model, and serving teams, with hybrid work options and relocation support for some

Qualifications

• Experience designing, building, and shipping developer tools or observability systems used by engineers.
• Fluency in systems/performance languages (C++, Rust, Go) and user-facing language (Python).
• Experience measuring software performance via profiling or benchmarking and analyzing results.

Responsibilities

• Enhance the MAX developer experience by creating tooling for traditional and agentic workflows.
• Design, build, and maintain profiling, tracing, debuggers, and dashboards for MAX deployments.
• Create agent-native tooling and write agents to manage models and triage issues.
• Integrate MAX tools with Mojo and driver APIs across orchestration layers.
• Participate in design discussions and code reviews to uphold high engineering standards.
• Collaborate with kernel, model, and serving engineers to turn pain points into better tools.

Skills

C++

Rust

Go

Python

Mojo familiarity

Tools

Nsight Systems

Perfetto

CUDA

PyTorch

Job description

Modular is seeking an engineer for the MAX Tools team to build developer tooling for its AI inference stack, enabling engineers to create, debug, and optimize models and kernels. You will work on profiling, tracing, debuggers, and dashboards, as well as agent-native tooling and integration across the stack.

The role emphasizes shipping developer tools, performance measurement, and close collaboration with kernel, model, and serving teams, with hybrid work options and relocation support for some