aipythonllmpytorchsftlorahuggingfacedpoqloragrpogpu training
Вакансия из Hirify Global, списка международных tech-компаний Для мэтча и отклика нужен Plus
Пожаловаться
Зарплата и рынок
развернуть
€75к-125к/год
€80к€110кмедиана€140к
По похожим ролям (46) ~ €110к/год
На СНГ рынке: $4.4к/мес ($3.7к - $5.9к)
85
Крутая вакансия
развернуть
Это четко описанная инженерная роль с выходом в production и масштабом в десятки миллионов пользователей; главный нюанс - работа с откровенным и потенциально вредным контентом.
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
LLM Engineer — Post-Training (LLM/AI): Fine-tuning self-hosted open-weight language models and shipping improved model behavior to a 60M-user AI companionship platform with an accent on SFT, LoRA/QLoRA, DPO, GRPO, data, and evaluation. Focus on diagnosing failures, building reproducible training loops, improving personalization and multilingual behavior, and taking experiments into production.
Location: Remote — Germany or Portugal; the role is also described as fully remote and work from anywhere.
Salary: €75K–€125K per year.
Company
hirify.global builds AI companionship products, including a large-scale AI companionship platform governed by its proprietary EverGuard moderation system.
What you will do
Fine-tune self-hosted, open-weight LLMs using SFT, LoRA/QLoRA, DPO, GRPO, and other post-training methods.
Diagnose model failures and determine whether to address them through data, post-training, prompting, or classifiers.
Build training data and evaluations that guide subsequent model improvements.
Improve personalization, character consistency, tone, multilingual behavior, and group interactions.
Ship models to production and use real user behavior to guide iteration.
Make training experiments faster, easier to compare, and more reproducible.
Requirements
Personally post-trained open-weight LLMs for a user-facing product and shipped them to production.
Experience working where AI was core to the product or business; fine-tuning must be central to the experience.
Strong Python, PyTorch, and Hugging Face skills, with hands-on GPU training and debugging.
Ability to build data and evaluations independently and make evidence-based decisions.
Ability to take work beyond notebooks into production.
Use AI tools daily in engineering work.
Fully comfortable working with uncensored models and explicit NSFW content, including harmful material, when developing and evaluating moderation systems.
Nice to have
Experience with conversational AI, roleplay products, or highly personalized consumer AI.
Experience with multilingual LLMs, moderation, safety, or adversarial testing.
Experience with multimodal models.
Open-source LLM or post-training contributions.
Culture & Benefits
Fully remote team with a culture of ownership, velocity, and creativity; annual in-person gathering.
B2B contract is preferred, with flexibility for long-term commitment.
20 working days of paid time off per year.
Monthly allowance of $100 for health insurance expenses, plus OpenUp access for the employee and up to three family members.
Co-working budget of up to twice per month, learning budget, company laptop, and equipment allowances.
Premium access to ChatGPT, Hugging Face, Claude Code, and other AI tools.
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 →