About the role
The role
We're looking for a Compute Software Engineer to build and own our virtualization layer—a critical component of Blaxel's AI-native cloud infrastructure. You'll work on our fork of Firecracker, the open-source microVM technology, extending and optimizing it to power the fastest serverless compute platform for agentic AI workloads.
This is a foundational role where you'll architect and implement core virtualization primitives that enable sub-25ms cold starts, secure multi-tenant isolation, and efficient resource management at scale.
What you'll do
-
Design and build low-level virtualization features on our Firecracker fork, including VM lifecycle management, memory optimization, and I/O performance improvements
-
Develop kernel-level optimizations and custom device models to minimize latency and maximize throughput for AI workloads
-
Build monitoring, observability, and debugging tools for production virtualization infrastructure
-
Work closely with the platform team to integrate the virtualization layer with our control plane and orchestration systems
-
Contribute upstream to Firecracker and engage with the open-source community
-
Profile and optimize critical paths in the VM creation and execution pipeline
-
Handle production incidents and drive continuous reliability improvements
Who you are
You're a systems engineer who thrives on building high-performance infrastructure from first principles. You care deeply about latency, efficiency, and correctness. You're comfortable reading kernel code, debugging at the hardware level, and making architectural tradeoffs that impact the entire stack.
