Member of Technical Staff - RL Training Framework

SpaceX was founded under the belief that a future where humanity is out exploring the stars is fundamentally more exciting than one where we are not. Today, SpaceX is actively developing the technologies to make this possible, with the ultimate goal of making life multiplanetary and expanding consciousness to the stars. SpaceXSI is building Super Intelligence to help humanity understand the universe. Our teams are building the full Super Intelligence stack: from our Colossus supercomputers, to Grok models, to our suite of applications, including Grok Bot, Grok Build, Grok Imagine, Grok Voice, and X. We're creating a powerful flywheel where more compute trains better models, better models power more useful products, and what we learn from people using those products shapes what we build next. ABOUT THE ROLE: The RL infrastructure team is looking for an engineer to help with low precision RL training and inference. RESPONSIBILITIES: Design and optimize our inference stack for all shapes of RL workloads at SpaceXSI, from small scale ablations to production training runs. Analyze, profile and address performance bottlenecks in large scale RL systems Work closely with the modelling team to efficiently implement novel RL techniques and algorithms BASIC QUALIFICATIONS: Experience in building, debugging, and optimizing efficiency of large-scale distributed systems Experience in LLM inference Proficiency in programming languages such as Python, C++ and/or Rust; frameworks such as PyTorch, Jax, CUDA Willingness to dive deep and solve hardcore problems at all levels of the stack PREFERRED SKILLS AND EXPERIENCE: Strong knowledge in quantization and numerics in LLM inference and training Experience in developing inference engines, e.g. SGLang, vLLM COMPENSATION AND BENEFITS: $180,000 - $440,000 USD Base salary is just one part of our total rewards package at SpaceXSI, which also includes equity, comprehensive medical, vision, and dental coverage, access to a 401(k) retirement pl

SpaceX was founded under the belief that a future where humanity is out exploring the stars is fundamentally more exciting than one where we are not. Today, SpaceX is actively developing the technologies to make this possible, with the ultimate goal of making life multiplanetary and expanding consciousness to the stars. SpaceXSI is building Super Intelligence to help humanity understand the universe. Our teams are building the full Super Intelligence stack: from our Colossus supercomputers, to Grok models, to our suite of applications, including Grok Bot, Grok Build, Grok Imagine, Grok Voice, and X. We're creating a powerful flywheel where more compute trains better models, better models power more useful products, and what we learn from people using those products shapes what we build next. ABOUT THE ROLE: As a multimodal engineer on the Imagine Model Team, you will develop cutting-edge AI experiences beyond text, with a strong focus on enabling high-fidelity understanding and generation across image and video modalities, while also incorporating audio where it enhances visual content (e.g., synchronized audio for video). Responsibilities span data curation, modeling, training, inference serving, and product integration, covering both pretraining and post-training phases. You will collaborate closely with product teams to push model frontiers and deliver exceptional end-to-end user experiences. RESPONSIBILITIES: Create and drive engineering agendas to advance multimodal capabilities, with emphasis on image and video generation, editing, understanding, controllable/long-horizon synthesis, agentic planning, RL training, and world simulation (including audio integration for richer video experiences). Improve data quality through annotation, filtering, augmentation, synthetic generation, captioning, and in-depth data studies, particularly for visual and audio data. Design evaluation frameworks, metrics, benchmarks, evals, and reward models tailored to image/video

SpaceX was founded under the belief that a future where humanity is out exploring the stars is fundamentally more exciting than one where we are not. Today, SpaceX is actively developing the technologies to make this possible, with the ultimate goal of making life multiplanetary and expanding consciousness to the stars. SpaceXSI is building Super Intelligence to help humanity understand the universe. Our teams are building the full Super Intelligence stack: from our Colossus supercomputers, to Grok models, to our suite of applications, including Grok Bot, Grok Build, Grok Imagine, Grok Voice, and X. We're creating a powerful flywheel where more compute trains better models, better models power more useful products, and what we learn from people using those products shapes what we build next. RESPONSIBILITIES: Scale synthetic coding data to trillions of tokens with large-scale docker verification. Distill the intelligence of flagship models into flash models through synthetic data generation. Optimize mid-training data mixtures to boost the ceiling for RL. Engineer long-context data recipes. Develop robust and diverse evaluation for mid-training checkpoints. BASIC QUALIFICATIONS: Expertise in ML and large model scaling, with familiarity across all kinds of scaling laws. Strong ability to design ML experiments. Familiarity with state-of-the-art techniques for curating AI training data for text, image, audio, and video modalities. Strong engineering abilities in Spark, Ray, and other frameworks for large-scale data processing. COMPENSATION AND BENEFITS: $180,000 - $440,000 USD Base salary is just one part of our total rewards package at SpaceXSI, which also includes equity, comprehensive medical, vision, and dental coverage, access to a 401(k) retirement plan, short & long-term disability insurance, life insurance, and various other discounts and perks. SpaceXSI is an equal opportunity employer. For details on data processing, view our Recruitment Privacy Not

SpaceX was founded under the belief that a future where humanity is out exploring the stars is fundamentally more exciting than one where we are not. Today, SpaceX is actively developing the technologies to make this possible, with the ultimate goal of making life multiplanetary and expanding consciousness to the stars. SpaceXSI is building Super Intelligence to help humanity understand the universe. Our teams are building the full Super Intelligence stack: from our Colossus supercomputers, to Grok models, to our suite of applications, including Grok Bot, Grok Build, Grok Imagine, Grok Voice, and X. We're creating a powerful flywheel where more compute trains better models, better models power more useful products, and what we learn from people using those products shapes what we build next. ABOUT THE ROLE: You will work on the most critical post-training and reinforcement learning challenges at any given time — including reward modeling, preference optimization (RLHF/DPO), and RL for improving reasoning, truthfulness, and real-world capabilities. You will get clarity on your first project before an offer. BASIC QUALIFICATIONS: You believe truth-seeking AI is the most important and challenging problem. You are obsessed about building incredibly useful models through post-training and RL techniques. You are a power user of AI models and eager to push the boundaries of what’s possible with reinforcement learning and alignment methods. If you previously worked on post-training, RLHF, or trained models used by millions of people it’s a big plus, but relevant experience is not required. You take pride in your work and thrive in meritocratic environments. COMPENSATION AND BENEFITS: $180,000 - $600,000 USD Base salary is just one part of our total rewards package at SpaceXSI, which also includes equity, comprehensive medical, vision, and dental coverage, access to a 401(k) retirement plan, short & long-term disability insurance, life insurance, and various other discou

SpaceX was founded under the belief that a future where humanity is out exploring the stars is fundamentally more exciting than one where we are not. Today, SpaceX is actively developing the technologies to make this possible, with the ultimate goal of making life multiplanetary and expanding consciousness to the stars. SpaceXSI is building Super Intelligence to help humanity understand the universe. Our teams are building the full Super Intelligence stack: from our Colossus supercomputers, to Grok models, to our suite of applications, including Grok Bot, Grok Build, Grok Imagine, Grok Voice, and X. We're creating a powerful flywheel where more compute trains better models, better models power more useful products, and what we learn from people using those products shapes what we build next. ABOUT THE ROLE: The RL infrastructure team is looking for an engineer to help develop our evaluation infrastructure. RESPONSIBILITIES: Build highly reliable, efficient, and easy to use infrastructure that runs all our evaluations at scale Collaborate closely with modeling & evaluation teams on developing new evals, maintaining evaluation signal quality, and building internal tooling to support training our next-generation of models Identify and resolve performance bottlenecks in all layers of the eval infrastructure stack, including inference, capacity fleet management, asynchronous eval orchestration, and more BASIC QUALIFICATIONS: Experience in building, debugging, and optimizing efficiency of large-scale distributed systems Willingness to dive deep and solve hard problems at all levels of the stack PREFERRED SKILLS AND EXPERIENCE: Experience in LLM inference Experience in GPU compute management, workload scheduling, and dynamic resource optimization Experience in developing interfaces for comparing models and evals COMPENSATION AND BENEFITS: $180,000 - $440,000 USD Base salary is just one part of our total rewards package at SpaceXSI, which also includes equity, comprehen

SpaceX was founded under the belief that a future where humanity is out exploring the stars is fundamentally more exciting than one where we are not. Today, SpaceX is actively developing the technologies to make this possible, with the ultimate goal of making life multiplanetary and expanding consciousness to the stars. SpaceXSI is building Super Intelligence to help humanity understand the universe. Our teams are building the full Super Intelligence stack: from our Colossus supercomputers, to Grok models, to our suite of applications, including Grok Bot, Grok Build, Grok Imagine, Grok Voice, and X. We're creating a powerful flywheel where more compute trains better models, better models power more useful products, and what we learn from people using those products shapes what we build next. ABOUT THE ROLE: You will join the multimodal team to push toward superhuman multimodal intelligence. Advance understanding and generation across modalities—image, video, audio, and text—spanning the full stack: data curation/acquisition, tokenizer training, large-scale pre-training, post-training/alignment, infrastructure/scaling, evaluation, tooling/demos, and end-to-end product experiences. Collaborate cross-functionally with pre-training, post-training, reasoning, data, applied, and product teams to deliver frontier capabilities in multimodal reasoning, world modeling, tool use, agentic behaviors, and interactive human-AI collaboration. Contribute to building models that can see, hear, reason about, and interact with the world in real time at unprecedented levels. RESPONSIBILITIES: Design, build, and optimize large-scale distributed systems for multimodal pre-training, post-training, inference, data processing, and tokenization at web/petabyte scale. Develop high-throughput pipelines for data acquisition, preprocessing, filtering, generation, decoding, loading, crawling, visualization, and management (images, videos, audio + text). Advance multimodal capabilities including sp