aipythonllmpytorchreinforcement learningsynthetic datasupervised fine tuningagent training
Вакансия из Hirify Global, списка международных tech-компаний Для мэтча и отклика нужен Plus
Пожаловаться
Зарплата и рынок
развернуть
На похожих ролях по рынку
$15к$45кмедиана$70к
9 в выборке
75
Хорошая вакансия
развернуть
Роль предлагает практическое исследование в растущей AI-компании с хорошим диапазоном зарплаты, но требование работать на месте может ограничить некоторых потенциальных кандидатов.
Требование на месте
Хороший диапазон зарплаты
Продуктовая компания
Исследовательский фокус
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
Research Intern (Agent RL Training) (AI/LLM): Exploring how large language models can support content understanding, recommendation, agentic web browsing, and autonomous multi-step task completion with an accent on SFT experiments, RL-based post-training, and agent datasets. Focus on designing end-to-end experiments, developing reward and training iterations, and building high-quality synthetic and preference data.
Location: Onsite in the Mountain View, California office, United States
Hourly pay: $35–$50 USD per hour
Company
hirify.global is a content intelligence platform serving personalized local news and information through AI, recommendation systems, and adtech.
What you will do
Work with a full-time mentor to identify research directions for applying large language models to hirify.global products.
Run end-to-end supervised fine-tuning experiments for LLM-based agents.
Explore reinforcement-learning post-training, including reward design and training iterations.
Build instruction-following, preference-pair, agent-trajectory, and synthetic training datasets.
Contribute to public research publications and top-venue submissions.
Requirements
Independent ability to run end-to-end model SFT experiments.
Basic understanding of RL-based post-training methods, including RLHF, DPO, PPO, and GRPO.
Strong Python and PyTorch skills.
Ability to evaluate model behavior across user-facing domains and explain quality criteria.
Strong motivation, research interest, intellectual curiosity, and willingness to work additional hours when needed.
Nice to have
Publication at a top-tier venue such as NeurIPS, ICML, ICLR, ACL, or EMNLP.
Experience with multi-node distributed training using FSDP, DeepSpeed, or Megatron-LM.
Experience writing custom GPU kernels with Triton or CUDA.
Experience building synthetic data pipelines for agent training.
Familiarity with TRL, OpenRLHF, or veRL/vLLM.
Culture & Benefits
Mentorship from a full-time employee throughout the internship.
Hands-on, from-zero-to-one research on LLM applications and agent capabilities.
Support for public research and top-venue publication submissions.
Potential eligibility for discretionary bonuses and options depending on the position.
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 →