{"schemaVersion":"jobsearcher.job.v1","id":"9c7cfc797cbfa06b8d4e8f8b","url":"https://jobsearcher.com/jobs/9c7cfc797cbfa06b8d4e8f8b","canonicalUrl":"https://jobsearcher.com/jobs/9c7cfc797cbfa06b8d4e8f8b","title":"Software Engineer (Full-Stack)","description":"AboutThis role is about making models better after their initial training. You'll work on reinforcement learning and other post-training methods, while also improving the systems needed to run those experiments efficiently at scale.\r\nThe company is building AI systems that learn how to carry out complex work inside large organisations. They recreate real-world workflows as interactive training environments, then use those environments to train models through practice and feedback — so the models get better at completing long, multi-step tasks reliably, rather than simply generating answers.\r\nYou'll work across both the learning algorithms and the infrastructure underneath them. That means going from an RL experiment to a GPU profiler trace, finding what's limiting performance, and making sure systems improvements don't change the way the model learns.\r\nWhat you'll doBuild and improve supervised fine-tuning, preference optimisation and RL methods\r\nWork with approaches including PPO, GRPO and SDPO\r\nOwn training loops from rollout generation through to policy updates and checkpointing\r\nImprove training throughput, GPU utilisation and memory efficiency\r\nProfile and fix bottlenecks across distributed training\r\nInvestigate instability and differences between training and inference\r\nUse real model failures to improve rewards, training data and overall performance\r\nWhat you'll needStrong programming and quantitative problem-solving skills\r\nHands-on experience with PyTorch and model training\r\nUnderstanding of reinforcement learning or LLM post-training\r\nExperience with distributed training and GPU systems\r\nStrong experimental judgement\r\nAbility to work comfortably across ML research and systems engineering\r\nOptionalvLLM, SGLang, Ray, FSDP or Megatron-LM\r\nCUDA, Triton, CuTE or GPU performance optimisation\r\nShortlisted candidates will be contacted within 48 hours.#J-18808-Ljbffr","company":"Axioma Search","rawCompany":"axioma search","city":"Millbrae","state":"CA","isRemote":false,"isActive":true,"createdAt":"2026-11-04T01:55:59.166Z","occupations":[{"code":"15-1252.00","title":"Software Developers","slug":"software-developers"},{"code":"15-1299.08","title":"Computer Systems Engineers/Architects","slug":"computer-systems-engineers-architects"},{"code":"15-1251.00","title":"Computer Programmers","slug":"computer-programmers"}],"industries":[{"code":"541511","title":"Custom Computer Programming Services","slug":"custom-computer-programming-services"},{"code":"513210","title":"Software Publishers","slug":"software-publishers"},{"code":"541512","title":"Computer Systems Design Services","slug":"computer-systems-design-services"}],"jobPosting":{"@context":"https://schema.org","@type":"JobPosting","title":"Software Engineer (Full-Stack)","description":"AboutThis role is about making models better after their initial training. You'll work on reinforcement learning and other post-training methods, while also improving the systems needed to run those experiments efficiently at scale.\r\nThe company is building AI systems that learn how to carry out complex work inside large organisations. They recreate real-world workflows as interactive training environments, then use those environments to train models through practice and feedback — so the models get better at completing long, multi-step tasks reliably, rather than simply generating answers.\r\nYou'll work across both the learning algorithms and the infrastructure underneath them. That means going from an RL experiment to a GPU profiler trace, finding what's limiting performance, and making sure systems improvements don't change the way the model learns.\r\nWhat you'll doBuild and improve supervised fine-tuning, preference optimisation and RL methods\r\nWork with approaches including PPO, GRPO and SDPO\r\nOwn training loops from rollout generation through to policy updates and checkpointing\r\nImprove training throughput, GPU utilisation and memory efficiency\r\nProfile and fix bottlenecks across distributed training\r\nInvestigate instability and differences between training and inference\r\nUse real model failures to improve rewards, training data and overall performance\r\nWhat you'll needStrong programming and quantitative problem-solving skills\r\nHands-on experience with PyTorch and model training\r\nUnderstanding of reinforcement learning or LLM post-training\r\nExperience with distributed training and GPU systems\r\nStrong experimental judgement\r\nAbility to work comfortably across ML research and systems engineering\r\nOptionalvLLM, SGLang, Ray, FSDP or Megatron-LM\r\nCUDA, Triton, CuTE or GPU performance optimisation\r\nShortlisted candidates will be contacted within 48 hours.#J-18808-Ljbffr","datePosted":"2026-11-04T01:55:59.166Z","dateModified":"2026-11-04T01:55:59.166Z","hiringOrganization":{"@type":"Organization","name":"Axioma Search","sameAs":"https://jobsearcher.com"},"jobLocation":{"@type":"Place","address":{"@type":"PostalAddress","addressLocality":"Millbrae","addressRegion":"CA","addressCountry":"US"}},"identifier":{"@type":"PropertyValue","name":"JobSearcher","value":"9c7cfc797cbfa06b8d4e8f8b"},"url":"https://jobsearcher.com/jobs/9c7cfc797cbfa06b8d4e8f8b"}}