Jobs at Modular

Build and operate a multi-cloud, multi-tenant platform for distributed LLM inference. Responsibilities include developing fault-tolerant services across cloud and hardware environments, improving observability and cold-start performance, and creating Helm charts and Kubernetes operators for scalable deployments. The role requires backend engineering, distributed systems, cloud infrastructure, Kubernetes, and machine learning experience.

Develop and mature the LLDB-based Mojo debugger, including interactive debugging, IDE integration, debug information, expression evaluation, GPU debugging, profiling, and MCP capabilities. Define the debugger roadmap, collaborate with compiler and developer relations teams, engage the developer community, and support open-source contributions. The role requires LLDB internals expertise and strong knowledge of debuggers, linkers, compilers, and systems programming.

Own Modular’s workplace IT systems, including Okta identity and access, onboarding and offboarding, SaaS applications, and an AI-powered help desk agent. Handle escalations, automate recurring support issues, administer SSO and SCIM provisioning, manage application licenses and vendors, and maintain documentation. The role combines hands-on IT operations with platform engineering, roadmap ownership, workflow development, and cross-functional collaboration with Security and AI Enablement.

Lead and grow a team of approximately 12 kernel engineers developing high-performance inference kernels across NVIDIA, AMD, and ASIC hardware. Own the MAX kernel roadmap, performance optimization, portable kernel architecture, and delivery. Partner with framework, serving, compiler/runtime, hardware enablement, and Qualcomm teams while advancing AI-assisted development, verification, and engineering practices.

Design, implement, and optimize high-performance attention and foundational AI kernels for GPUs, NPUs, ASICs, and other accelerators. Build reusable attention-framework abstractions, profile workloads, develop benchmarks and correctness tests, analyze hardware and compiler behavior, and improve production inference performance. Collaborate with hardware and software teams, document designs, resolve ambiguous performance issues, and mentor engineers while supporting new accelerator architectures and AI workloads.

Own developer-facing product marketing for Mojo and MAX, including release messaging, technical content, launch plans, executive presentations, community engagement, competitive positioning, and adoption analytics. Partner with engineering, product management, and developer advocates to explain AI infrastructure, GPU performance, inference optimization, and programming language capabilities to technical audiences.

Design and build modular, high-performance distributed systems for AI model inference. Lead architecture decisions involving concurrency, data movement, APIs, system boundaries, and reliability. Develop extensible interfaces and composable components for model developers and contributors. Diagnose complex performance bottlenecks and support production software. The role values systems programming, abstraction design, GPU and accelerator optimization, inference frameworks, and research awareness.

The role involves optimizing hardware support within a software stack for AI model deployment, collaborating with teams and hardware vendors to enhance performance and portability across platforms.

Design and implement Mojo language features, compiler optimizations, and multi-target compilation for CPUs, GPUs, and AI accelerators using MLIR and LLVM. Collaborate with kernel authors, hardware engineers, researchers, and the open-source community to improve language design, code generation, performance, developer experience, and the broader Mojo SDK.

Build and evolve Mojo developer tooling including an LLDB-based debugger, IDE integrations, diagnostics, profiling, and GPU debugging. Partner with compiler engineers, shape roadmap, collaborate with DevRel and open-source community, and deliver production-quality debugging and developer workflows.

Lead and grow a compiler, kernel, and runtime engineering team responsible for accelerator bring-up across the Modular stack. Drive technical direction, architecture, coding, roadmap delivery, vendor partnerships, SOWs, milestones, cross-team dependencies, hiring, and production-quality model deployment across diverse hardware platforms.

As a Mojo Libraries Engineer, you'll design and implement libraries for Mojo, focusing on performance and integration with AI systems, collaborating with multiple teams to optimize algorithms and deliver high business impact.

Lead technical vision and architecture for the Mojo compiler backend: IR lowering, optimization passes, and code generation across CPUs, GPUs, and AI accelerators. Drive MLIR/LLVM initiatives, expand debugging and profiling tooling (LLDB/GDB, LSP, profilers), mentor engineers, collaborate with cross-functional teams, and engage with the open-source community to shape developer tooling and compiler strategy.

Design and develop AI runtimes and cross-stack optimizations for CPU, GPU, and accelerator platforms. Improve performance through caching, data locality, networking, profiling, and benchmarking. Collaborate with compiler, kernel, serving, model, infrastructure, tooling, and customer success teams to deliver high-performance AI systems and address customer requirements.

As a Senior AI Kernel Engineer, you will design and optimize high-performance kernels for AI inference on GPUs and custom accelerators, improving performance and collaborating with various teams to influence architecture and implementation.

Drive the product vision for Modular's cloud compute platform, collaborating with teams to enhance AI service capabilities and customer engagement.

Profile customer LLM inference workloads end-to-end and apply optimizations across GPU kernels, inference engine, and distributed systems. Build tooling and a repeatable automated optimization platform, partner with GTM and engineering to deploy tuned inference in production, and publish technical insights to shape industry best practices.

Design, build, and maintain developer tooling for the MAX AI inference stack, including profilers, tracers, debuggers, dashboards, and agent-native tools. Integrate tooling across MAX, Mojo, and driver APIs, collaborate with kernel/model/serving engineers, and drive performance analysis, debugging, and observability to improve developer productivity and production reliability.