Job description

About the Role RadixArk is looking for a Member of Technical Staff — TPU Systems to build high-performance inference and training systems using JAX, XLA, and Pallas. You'll push model workloads to their limits on TPU hardware, working on SGLang-JAX and other critical infrastructure that enables efficient deployment of frontier models on Google's tensor processing units. Requirements 3+ years experience building production ML systems utilizing JAX/Torch, XLA, or TPU-focused frameworks. Bachelor's or Master's degree in Computer Science, Electrical Engineering, or equivalent industry experience Deep understanding of XLA internals preferred: HLO, MLIR, operator fusion, SPMD partitioning, and sharding strategies. Strong performance tuning instincts across compiler and runtime layers Experience with distributed inference systems (e.g. SGLang, vLLM) or training frameworks (e.g. Miles, Alpa, Pathways) Proficiency in Python with demonstrated ability to write high-performance, production-quality code Experience writing custom GPU/TPU/AI Accelerator kernels. Familiarity with Pallas for kernel development is strongly preferred. Responsibilities Build high-performance inference and training systems using JAX/XLA/Pallas, including SGLang-JAX Push large-model workloads to the limits on the newest TPU hardwares Optimize end-to-end latency and throughput for LLM serving on TPU infrastructure Design and implement SPMD strategies for efficient distributed inference and training Design and implement Pallas kernels for operations that require customized low level control for best performance Profile and optimize XLA compilation pipelines and HLO graph transformations Collaborate with kernel engineers and compiler teams to achieve performance wins across the stack Contribute to open-source projects with TPU optimization guides, benchmarks, and architectural insights About RadixArk RadixArk is an infrastructure-first company built by engineers who've shipped production AI systems, created SGLang (30K+ GitHub stars, the fastest open LLM serving engine), and developed Miles (our large-scale RL framework). Founded by AI infrastructure veterans from xAI and NVIDIA, we're on a mission to democratize frontier-level AI infrastructure by building world-class open systems for inference and training. Our team has optimized kernels serving billions of tokens daily, designed distributed training systems coordinating 10,000+ GPUs, and contributed to infrastructure that powers leading AI companies and research labs. Compensation Depending on background, skills, and experience, the expected annual salary range for this position is $200,000 - $400,000 USD + equity. Equal Opportunity RadixArk is an Equal Opportunity Employer and is proud to offer equal employment opportunity to everyone regardless of race, color, ancestry, religion, sex, national origin, sexual orientation, age, citizenship, marital status, disability, gender identity, veteran status, and more.

Requirements

Must-haves

3+ years experience building production ML systems utilizing JAX/Torch, XLA, or TPU-focused frameworks.
Bachelor's or Master's degree in Computer Science, Electrical Engineering, or equivalent industry experience
Strong performance tuning instincts across compiler and runtime layers
Experience with distributed inference systems (e.g. SGLang, vLLM) or training frameworks (e.g. Miles, Alpa, Pathways)
Proficiency in Python with demonstrated ability to write high-performance, production-quality code
Build high-performance inference and training systems using JAX/XLA/Pallas, including SGLang-JAX
Push large-model workloads to the limits on the newest TPU hardwares
Optimize end-to-end latency and throughput for LLM serving on TPU infrastructure
Design and implement SPMD strategies for efficient distributed inference and training
Design and implement Pallas kernels for operations that require customized low level control for best performance
Profile and optimize XLA compilation pipelines and HLO graph transformations
Collaborate with kernel engineers and compiler teams to achieve performance wins across the stack
Contribute to open-source projects with TPU optimization guides, benchmarks, and architectural insights

Nice-to-haves

Deep understanding of XLA internals preferred: HLO, MLIR, operator fusion, SPMD partitioning, and sharding strategies.
Experience writing custom GPU/TPU/AI Accelerator kernels. Familiarity with Pallas for kernel development is strongly preferred.

Application prep

Use the role’s language and application details while tailoring your materials.

Keywords to mirror

mlpythonaillm

Platform: greenhouse
Effort: quick