Member of Technical Staff — Training

RadixArk; Palo Alto (CA); On-site; USD 200,000 - 400,000; Full time

Benefits offered by this job

401(k) plan

Unlimited PTO

Visa sponsorship

Job summary

RadixArk is seeking a Member of Technical Staff, Training, to design, build, and operate the distributed systems behind large-scale model post-training. You will work across training, inference, and orchestration on multi-GPU clusters to ensure performance, correctness, and reliability.

The role blends modeling, infrastructure, and low-level systems work, with opportunities to contribute to Miles, SGLang, and other open-source projects.

Qualifications

• 3+ years of experience building or operating distributed ML systems.
• Hands-on experience with post-training backends or inference systems for large language models (e.g., Megatron-LM, FSDP, SGLang, TensorRT-LLM, vLLM).
• Experience in at least two of the following areas: performance/scalability of multi-GPU, multi-node workloads; numerical correctness or low precision; stability or fault tolerance; post-training recipes and orchestration for large training runs; multimodal training or inference; agent infrastructure and sandboxed eval systems; open-source projects.

Responsibilities

• Design, build, and operate distributed training, rollout, and orchestration systems for large-scale LLM and multimodal post-training across multi-GPU, multi-node environments.
• Profile and optimize performance across the full-stack — model implementation, parallelism strategies, communication libraries, and GPU kernels.
• Investigate numerical correctness and low-precision issues in distributed training and inference, including train–inference consistency for reinforcement learning.
• Improve the reliability of long-running workloads through checkpointing, fault recovery, observability, and operational tooling.
• Build supporting infrastructure for reinforcement learning and agentic post-training, including asynchronous rollout, trajectory collection, sandboxed execution, evaluation harnesses, and data pipelines.
• Contribute to open-source training and inference systems, including Miles and SGLang, and partner with researchers to turn experimental requirements into production systems.

Skills

Distributed ML systems

Multi-GPU training

High-performance inference

Tools

Miles

SGLang

Megatron-LM

TensorRT-LLM

vLLM

NCCL

RDMA

NVLink/NVSwitch

CUDA

Triton

CUTLASS

Job description

About the Role

As a Member of Technical Staff, Training, you will design, build, and operate the distributed systems behind large-scale model post-training — spanning training, inference, and orchestration, with a focus on the performance, correctness, scalability, and reliability of workloads running across large GPU clusters.

This role suits engineers who move fluidly across modeling recipes, complex infrastructure, and low-level systems, identify bottlenecks in distributed workloads, and translate experimental requirements into robust software.

In This Role, You Will

• Design, build, and operate distributed training, rollout, and orchestration systems for large-scale LLM and multimodal post-training across multi-GPU, multi-node environments.

• Profile and optimize performance across the full-stack — model implementation, parallelism strategies, communication libraries, and GPU kernels — to improve throughput, latency, memory efficiency, hardware utilization, and cost.

• Investigate numerical correctness and low-precision issues in distributed training and inference, including train–inference consistency for reinforcement learning.

• Improve the reliability of long-running workloads through checkpointing, fault recovery, observability, and operational tooling.

• Build supporting infrastructure for reinforcement learning and agentic post-training, including asynchronous rollout, trajectory collection, sandboxed execution, evaluation harnesses, and data pipelines.

• Contribute to open-source training and inference systems, including Miles and SGLang, and partner with researchers to turn experimental requirements into production systems.

Minimum Qualifications

• 3+ years of experience building or operating distributed machine learning systems, large-scale training infrastructure, or high-performance inference systems.

• Hands‑on experience with post‑training systems, training backends, or inference systems for large language models (e.g., Megatron‑LM, FSDP, SGLang, TensorRT‑LLM, vLLM).

• Experience in at least two of the following areas:

  • Performance, efficiency, and scalability of multi‑GPU, multi‑node workloads

  • Numerical correctness or low precision

  • Stability, reliability, or fault tolerance

  • Post‑training algorithm recipes and orchestration infrastructure for large training runs

  • Multimodal training or inference, including vision‑language models and multimodal generation

  • Agent infrastructure, including sandboxes, harnesses, and eval systems

  • Building and maintaining open‑source projects widely adopted in industry and academia

Preferred Qualifications

• Familiarity with RL algorithms such as PPO, GRPO, and their variants, and experience applying them in large-scale post‑training.

• Experience with modern post‑training frameworks (e.g., Miles, slime, AReaL, verl, Prime‑RL).

• Key open‑source contributions to training or inference frameworks (e.g., SGLang, vLLM, Megatron‑LM).

• GPU kernel development (e.g., CUDA, Triton, CUTLASS) or communication‑layer optimization (e.g., NCCL, RDMA, NVLink/NVSwitch).

• Experience training or serving models at very large scale (e.g., Mixture‑of‑Experts models on clusters of thousands of GPUs).

• Top‑tier publications in ML systems or other systems fields.

Even if you don't meet every qualification above, we encourage you to apply — we care most about demonstrated ability to build and reason about large‑scale systems.

About RadixArk

RadixArk builds open‑source and production infrastructure for large language models and multimodal post‑training. Our systems — including Miles, an enterprise‑grade reinforcement learning training framework, and SGLang, a widely deployed high‑performance LLM inference engine — power distributed post‑training across clusters of 10k–100k+ GPUs.

Compensation

Depending on background, skills, and experience, the expected annual salary range for this position is $200,000 to $400,000, plus equity.

Benefits include a 401(k) plan and unlimited PTO.

RadixArk sponsors employment visas (e.g., H-1B, O-1) for eligible candidates.

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.