Platform Engineer — Serverless AI (Remote — Worldwide)

Join the team building the foundational infrastructure for the next generation of AI applications. As a Platform Engineer at Modal, you will design and optimize a high-performance serverless runtime that enables developers to run data-intensive workloads, from LLM inference to large-scale media processing, without managing infrastructure. You will work at the intersection of container orchestration, Linux internals, and distributed systems to ensure our users can scale from zero to thousands of GPUs in seconds.

- Design and implement core features for the Modal runtime to minimize cold-start latency for containerized AI workloads. - Optimize GPU resource allocation and scheduling across heterogeneous clusters of cloud providers. - Enhance and maintain our custom container execution engine to support massive concurrency and high-throughput data I/O. - Build robust telemetry and observability tools to monitor platform health and performance at scale. - Collaborate with the product team to design APIs that simplify complex infrastructure patterns for AI engineers. - Participate in a follow-the-sun on-call rotation to ensure the stability of our global serverless fleet.

- Deep expertise in systems programming, preferably with Rust, C++, or Go. - Strong understanding of Linux internals, including namespaces, cgroups, and storage drivers. - Experience building or operating high-scale distributed systems or container runtimes (containerd, Firecracker, or gVisor). - Familiarity with the NVIDIA stack (CUDA, nvidia-container-toolkit) and GPU resource management. - Ability to work independently in a remote-first, fast-paced startup environment. - Proven track record of shipping complex infrastructure projects that prioritize low latency and high reliability.

Equity, healthcare, learning budget.