Research Engineer, Scaling

Target start date: Immediately. Relocation provided.

Since its founding in 2015, 1X has been at the forefront of developing advanced humanoid robots designed for household use. Our mission is to create an abundant supply of labor via safe, intelligent humanoids. At 1X, you'll own critical projects, tackle unsolved research problems, deliver great products to customers, and be rewarded based on merit and achievement.

As a Research Engineer, Scaling, you'll build the systems that let every team and every robot go faster: training more often, evaluating more reliably, and deploying better models to our growing fleet. You'll transform prototypes into production-scale infrastructure for learning and inference, enabling larger training runs and maximizing edge compute utilization to make our models more capable.

Python / C++

Triton / CUDA

Location

The role is based in Palo Alto, CA. Candidates are expected to be in-person at the office.

Responsibilities

• High agency and ownership on scaling capabilities in distributed training and/or inference
• Ensure that compute is never the bottleneck, i.e. we always have more compute available than data
• Enable large-scale (1000+ GPU) training on billion frames+ of robot data, from fault tolerance to distributed ops to experiment management
• Optimize high-throughput datacenter scale distributed inference for world models: work on the world's fastest diffusion inference engine
• Improve low-latency on-device inference for a variety of robot policies with quantization, scheduling, distillation and more

Requirements

• You must be scaling-pilled, and believe that scale will enable humanoid robots to exist, and be excited about being on the team that will make that happen for the first time in human history
• Python and/or C++ programming experience
• An intuitive understanding of training or inference scaling and what makes models run fast or slow

Ideal Experiences

• Degree in Computer Science or a related field
• Hands-on experience with distributed training (TorchTitan/Accelerate/DeepSpeed, FSDP/ZeRO, NCCL), multi-node debugging, and experiment management
• Depth in inference performance: TensorRT or similar graph compilers, batching/scheduling, and serving systems
• Real familiarity with quantization (PTQ, QAT; calibration strategies; INT8/FP8; libraries such as TensorRT ModelOpt, bitsandbytes, or equivalent)
• Experience writing or tuning CUDA/Triton kernels and leveraging vectorization, tensor cores, and memory hierarchy

Sample Projects

• Quantizing, pruning, distilling, and optimizing a model to run as fast as possible on a given hardware SKU
• Creating or contributing to large-scale, high-throughput inference engines for diffusion models or LLMs, like xDiT, SGLang, vLLM, or TensorRT-LLM
• Training large models on hundreds to thousands of GPUs, and designing the infrastructure to scale small experiments to large production runs

Interview Process

• The team reviews your CV and statement of exceptional work
• 15 minute phone conversation with our talent acquisition team
• 45-minute virtual interview with a team member asking a coding question in the language of your choice
• On-site interview (in-person or virtual) consisting of 4 technical interviews (mix of coding, systems design, open-ended research interview)
• Background reference checks
• Offer

Compensation

At 1X your work and results will be rewarded with a total rewards package consisting of a base salary, stock options and benefits. Base salary range is $180,000 to $300,000. Your actual salary will be based on your knowledge, skills and experience.