Member of Technical Staff — Diffusion Model

<h2 data-start="330" data-end="351"><strong data-start="333" data-end="351">About the Role</strong></h2> <p data-start="353" data-end="472">RadixArk is seeking a<span class="Apple-converted-space">&nbsp;</span><strong data-start="375" data-end="422">Member of Technical Staff — Diffusion Model</strong><span class="Apple-converted-space">&nbsp;</span>to advance the frontier of generative modeling.</p> <p data-start="474" data-end="787">You will work on cutting-edge diffusion and flow-based models for image, video, and multimodal generation, pushing model quality, efficiency, and scalability. This role combines deep research thinking with strong engineering execution — from designing novel algorithms to training and deploying models at scale.</p> <p data-start="789" data-end="918">Your work will directly shape next-generation generative AI systems used by researchers, developers, and real-world applications.</p> <p data-start="920" data-end="1054">This is a high-impact role for engineers and researchers who want to push the limits of generative models in both theory and practice.</p> <h2 data-start="1061" data-end="1080"><strong data-start="1064" data-end="1080">Requirements</strong></h2> <ul data-start="1082" data-end="1653"> <li data-start="1082" data-end="1149"> <p data-start="1084" data-end="1149">5+ years of experience in ML research or applied ML engineering</p> </li> <li data-start="1150" data-end="1263"> <p data-start="1152" data-end="1263">Strong expertise in diffusion models or generative models (DDPM, DDIM, latent diffusion, flow matching, etc.)</p> </li> <li data-start="1264" data-end="1333"> <p data-start="1266" data-end="1333">Deep understanding of deep learning fundamentals and optimization</p> </li> <li data-start="1334" data-end="1396"> <p data-start="1336" data-end="1396">Proven experience training large-scale models on GPUs/TPUs</p> </li> <li data-start="1397" data-end="1437"> <p data-start="1399" data-end="1437">Strong proficiency in PyTorch or JAX</p> </li> <li data-start="1438" data-end="1501"> <p data-start="1440" data-end="1501">Experience implementing research ideas into working systems</p> </li> <li data-start="1502" data-end="1581"> <p data-start="1504" data-end="1581">Strong mathematical foundation in probability, statistics, and optimization</p> </li> <li data-start="1582" data-end="1653"> <p data-start="1584" data-end="1653">Ability to move from research prototypes to production-quality models</p> </li> </ul> <h3 data-start="1660" data-end="1679"><strong data-start="1664" data-end="1679">Strong Plus</strong></h3> <ul data-start="1681" data-end="2122"> <li data-start="1681" data-end="1755"> <p data-start="1683" data-end="1755">Publications in top-tier conferences (NeurIPS, ICML, ICLR, CVPR, etc.)</p> </li> <li data-start="1756" data-end="1808"> <p data-start="1758" data-end="1808">Experience with large-scale distributed training</p> </li> <li data-start="1809" data-end="1878"> <p data-start="1811" data-end="1878">Experience in multimodal generation (text-to-image, video, audio)</p> </li> <li data-start="1879" data-end="1943"> <p data-start="1881" data-end="1943">Familiarity with transformer architectures and hybrid models</p> </li> <li data-start="1944" data-end="2009"> <p data-start="1946" data-end="2009">Experience improving sampling speed and generation efficiency</p> </li> <li data-start="2010" data-end="2068"> <p data-start="2012" data-end="2068">Contributions to open-source generative model projects</p> </li> <li data-start="2069" data-end="2122"> <p data-start="2071" data-end="2122">Experience scaling models to billions of parameters</p> </li> </ul> <h2 data-start="2129" data-end="2152"><strong data-start="2132" data-end="2152">Responsibilities</strong></h2> <ul data-start="2154" data-end="2672"> <li data-start="2154" data-end="2224"> <p data-start="2156" data-end="2224">Design and develop next-generation diffusion and generative models</p> </li> <li data-start="2225" data-end="2290"> <p data-start="2227" data-end="2290">Improve model quality, controllability, and sample efficiency</p> </li> <li data-start="2291" data-end="2353"> <p data-start="2293" data-end="2353">Research and implement novel training and sampling methods</p> </li> <li data-start="2354" data-end="2410"> <p data-start="2356" data-end="2410">Optimize models for large-scale distributed training</p> </li> <li data-start="2411" data-end="2477"> <p data-start="2413" data-end="2477">Collaborate with systems teams to scale training and inference</p> </li> <li data-start="2478" data-end="2540"> <p data-start="2480" data-end="2540">Translate research ideas into practical production systems</p> </li> <li data-start="2541" data-end="2598"> <p data-start="2543" data-end="2598">Evaluate models using rigorous metrics and benchmarks</p> </li> <li data-start="2599" data-end="2672"> <p data-start="2601" data-end="2672">Contribute to long-term research and product direction in generative AI</p> </li> </ul> <h2 data-start="2679" data-end="2700"><strong data-start="2682" data-end="2700">About RadixArk</strong></h2> <p>RadixArk is an infrastructure-first company built by engineers who've shipped production AI systems, created SGLang (20K+ GitHub stars, the fastest open LLM serving engine), and developed Miles (our large-scale RL framework).</p> <p>We're on a mission to democratize frontier-level AI infrastructure by building world-class open systems for inference and training.</p> <p>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.</p> <p>We're backed by well-known infrastructure investors and partner with Nvidia, Google, AWS, and frontier AI labs.</p> <p>Join us in building infrastructure that gives real leverage back to the AI community.</p> <h2 data-start="3369" data-end="3388"><strong data-start="3372" data-end="3388">Compensation</strong></h2> <p data-start="3390" data-end="3560">We offer competitive compensation with meaningful equity, comprehensive benefits, and flexible work arrangements. Compensation depends on location, experience, and level.</p> <h2 data-start="3567" data-end="3591"><strong data-start="3570" data-end="3591">Equal Opportunity</strong></h2> <p data-start="3593" data-end="3680">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.</p>