Назад
Company hidden
1 день назад

ML Engineer, Model Development (AI)

Формат работы
onsite
Тип работы
fulltime
Грейд
senior
Английский
c1
Страна
France/Spain/Germany
Вакансия из списка Hirify.GlobalВакансия из Hirify Global, списка международных tech-компаний
Для мэтча и отклика нужен Plus

Мэтч & Сопровод

Для мэтча с этой вакансией нужен Plus

Описание вакансии

Текст:
/
TL;DR
ML Engineer, Model Development (AI): Post-training, evaluating and optimising open language and multimodal models for aerospace engineering on sovereign, on-premise infrastructure with an accent on reproducible pipelines, model quality, safety and inference efficiency. Focus on designing evaluation suites, applying fine-tuning and alignment methods, optimising models for production, and releasing documented, versioned models for engineering applications.

Location: Toulouse, Munich, Madrid, Paris, or Hamburg, France, Germany, or Spain

Company

hirify.global develops aerospace, defense, space and connected-service technologies, with a focus on sustainable aerospace and European security and sovereignty.

What you will do

  • Post-train open language and multimodal models for engineering tasks using supervised fine-tuning, preference alignment, reinforcement learning and distillation.
  • Build reproducible training and post-training pipelines that support model retraining and continuous improvement.
  • Design evaluation suites covering domain accuracy, reasoning, hallucination, robustness and safety, and use the results to determine production releases.
  • Optimise models for on-premise deployment through quantisation, pruning and inference tuning while balancing quality, latency and GPU cost.
  • Release tested, versioned models with model cards, known limitations and usage contracts, and hand them over to agent and engineering teams.
  • Translate engineering needs into model requirements and measure the impact of each release through accuracy, time-saved and cost metrics.

Requirements

  • Master’s degree in computer science, machine learning or a related field.
  • Several years of hands-on experience fine-tuning or training deep learning models and deploying them for real users.
  • Strong Python and PyTorch skills, with experience in the Hugging Face ecosystem, including Transformers, TRL or PEFT.
  • Experience building ML model evaluations with benchmarks, test sets and metrics that reflect real-world use.
  • Strong engineering practices covering version control, testing, experiment tracking and reproducibility.
  • Fluent English required.

Nice to have

  • Experience with LLM preference alignment or reinforcement learning, such as DPO, GRPO or RLHF.
  • Experience with distillation, quantisation or inference engines such as vLLM, TensorRT-LLM or llama.cpp.
  • Experience with LLM evaluation frameworks or LLM-as-judge methods.
  • Distributed training at scale using technologies such as FSDP, DeepSpeed or Megatron-LM.
  • Background in engineering, manufacturing or aerospace data.
  • German language skills.

Culture & Benefits

  • Work on open models and sovereign, on-premise infrastructure using real engineering data.
  • Opportunity to influence AI use in aircraft engineering and contribute to real products.
  • Startup-style drive combined with the expertise of an established technology site.
  • Professional development opportunities within the hirify.global group.
  • Permanent, full-time employment with flexible working arrangements where possible.
  • Inclusive workplace committed to diversity, equal opportunity and integrity.

Hiring process

  • Submit a CV or application; the submitted information may be stored and used for application or future-employment monitoring.
  • hirify.global does not request monetary exchanges during recruitment; suspected impersonation should be reported to the stated hirify.global contact.

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 →