aipytorchasicvllmreinforcement learningkerneldistributed trainingpost training
Вакансия из Hirify Global, списка международных tech-компаний Для мэтча и отклика нужен Plus
Пожаловаться
Зарплата и рынок
Выше рынка
$400к/год
$230к$250кмедиана$300к
По похожим ролям (49) ~ $250к/год
На СНГ рынке: $4.1к/мес ($3.4к - $6к)
93
Отличная вакансия
развернуть
Узкая роль на переднем крае AI сочетает зарплату заметно выше рыночного ориентира, самостоятельную работу над post-training и доступ к большим вычислительным ресурсам. Главный компромисс - обязательная работа в офисе в Сан-Хосе.
Кликните для подробной информации
Только офисПередовой стекЗарплата выше рынкаПередовой AIБольшая ответственность
Оценка от Hirify AI
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
Research Engineer (AI): Training proprietary frontier models to automate kernel development, ASIC engineering, production pipelines, and 3-D hardware operating systems with an accent on post-training, reinforcement learning, and verifiable rewards. Focus on building end-to-end training stacks, simulation-driven environments, custom data mixes, and benchmarks that improve engineering productivity without reward hacking.
Location: On-site in San Jose, United States. Fully in-person work at the Santana Row office; relocation support is available for those moving to San Jose.
Salary: $400K+ plus significant equity.
Company
hirify.global builds inference-focused AI hardware and software by co-designing chips, racks, software, and manufacturing systems for high throughput and low latency.
What you will do
Own the post-training stack end to end, including supervised fine-tuning and reinforcement learning.
Build reinforcement learning environments with ASIC, kernel, and supercomputing engineers.
Use simulations, formal tools, tests, and hardware measurements to create verifiable rewards.
Create custom data mixes and train proprietary frontier models for difficult engineering tasks.
Develop benchmarks that improve model productivity and prevent reward hacking.
Optimize reinforcement learning systems to make training loops faster and more cost-efficient.
Requirements
Hands-on post-training experience, including SFT and RL such as PPO or GRPO, with shipped models used by people.
Experience improving a model from poor to strong performance on a task it was not built for, ideally with scarce data.
Strong PyTorch and distributed-training engineering skills, including the ability to inspect rollouts and validate metrics.
Strong understanding of the systems side of reinforcement learning.
Ability to make effective engineering decisions under compute and data constraints.
Willingness to work fully in person in San Jose.
Nice to have
Experience with reinforcement learning or post-training frameworks such as verl, OpenRLHF, TRL, or prime-rl.
Experience with inference engines such as vLLM or SGLang.
Experience with agentic, multi-turn reinforcement learning with verifiable rewards.
Experience with distillation and training small specialized models for production.
Experience in ASIC design or verification, accelerator kernels, or large-scale compute infrastructure.
Former founder experience.
Culture & Benefits
Medical, dental, and vision packages with generous premium coverage.
$500 monthly credit for waiving medical benefits.
$2,500 monthly housing subsidy for employees living within walking distance of the office.
Wellness benefits covering fitness and mental health.
Daily lunch and dinner at the office.
Unlimited compute budget subject to return-on-investment justification, plus significant equity.
Будьте осторожны: если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого - это мошенники. Обязательно жмите "Пожаловаться" или пишите в поддержку. Подробнее в гайде →