- Страна
- Франция
Откликайтесь
на вакансии с ИИ

Senior Data Scientist NLP/GenAI - Catalog
Отличная вакансия для Senior-специалиста: работа с передовыми технологиями (GenAI, LLM), прозрачный процесс найма, сильная команда и реальное влияние продукта на мировой рынок e-commerce.
Сложность вакансии
Высокая сложность обусловлена требованиями к глубокому опыту в NLP и GenAI (4+ года), владением стеком Spark/Databricks и необходимостью прохождения многоэтапного отбора, включая тестовое задание.
Анализ зарплаты
Зарплата в объявлении не указана, но для позиции Senior Data Scientist в Париже рыночный диапазон составляет 65,000–85,000 евро в год. Mirakl, как успешная компания из списка Next40, обычно предлагает конкурентоспособные условия на уровне или выше медианы рынка.
Сопроводительное письмо
I am writing to express my strong interest in the Senior Data Scientist NLP/GenAI position at Mirakl. With over 4 years of experience in developing and deploying machine learning models, particularly in the realm of NLP and Transformers, I am excited by the opportunity to contribute to your Catalog team. My background in fine-tuning LLMs and working with large-scale data processing using Spark aligns perfectly with Mirakl's mission to optimize marketplace operations through cutting-edge AI.
In my previous roles, I have successfully moved multiple algorithms from prototype to production, focusing on business-driven outcomes. I am particularly impressed by Mirakl's pioneer spirit in being one of the few French companies with fine-tuned LLMs in large-scale production. I am eager to bring my expertise in Python, PyTorch, and LangChain to help solve complex challenges like automatic content rewriting and product categorization for your 450+ global marketplaces.
Составьте идеальное письмо к вакансии с ИИ-агентом

Откликнитесь в mirakllabs уже сейчас
Присоединяйтесь к лидеру рынка e-commerce и внедряйте передовые LLM-решения в промышленном масштабе!
Описание вакансии
About Mirakl:
Founded in 2012, Mirakl has been at the forefront of marketplace innovation, empowering every business to compete in the platform economy.
Today, Mirakl’s operating system combines an enterprise marketplace solution (Mirakl Platform) that enables retailers and B2B organizations to launch, scale, and operate marketplaces and dropship, AI-powered multichannel selling (Mirakl Connect), retail media (Mirakl Ads) and an agentic commerce infrastructure (Mirakl Nexus).
With dual headquarters in Boston and Paris, Mirakl helps a global ecosystem of 450+ marketplaces (B2C and B2B) and a network of over 100k third-party marketplace sellers. Brands like Macy’s, Decathlon, Carrefour, Asos, and Airbus Helicopters use Mirakl to grow their businesses in new and remarkable ways.
For more information: www.mirakl.com.
Mirakl in Numbers:
- 🗓️ Founded in 2012 | Member of French Tech Next40
- 👥 750+ employees in 9 offices worldwide: Paris, Barcelona, Bordeaux, Boston, London, Munich, New York, Sydney, Tokyo
- 🇫🇷 350+ Mirakl Tech teams members mainly based in France
- ⚙️ 5 Saas Solutions
Our Values:
Working at Mirakl means accelerating your career alongside ambitious, passionate, and supportive colleagues. We're proud of the diversity of backgrounds, perspectives, and experiences that make our teams unique.
Our 5 values guide how we collaborate:
- 💡 Work Hard Together: Teamwork and collaboration are the foundation of our success
- 🏆 Get Things Done: We prioritize action and efficiency for impactful results
- 🚀 Go Above & Beyond: We tackle challenges proactively and always aim for excellence
- 🎓 Succeed Through Expertise: Knowledge sharing and continuous learning are core to our culture
- 🤝 Satisfy & Empower Clients: We're committed to our clients' success
About the job
You’ll join our Data Science team, where your main mission will be to prototype, iterate, and ship algorithms to production in close collaboration with Product, Data Engineering, and Software teams. Your projects will focus on Marketplace catalog challenges, including NLP, Computer Vision, and large-scale Generative AI (custom LLMs). The topics you’ll tackle will have a real impact on our customers: we aim to make the most of our rich, diverse data to grow their revenue, streamline marketplace operations, and ensure user and transaction safety.
As for remote set-up it would be:
- 4 days worked from our offices per week
- A day worked remotely per week
We’re hiring on a permanent contract (CDI), based in our Paris or Bordeaux Office (1 day remote per week). As part of our Data team (60+ people), you will work on:**
Catalog topics:
- Automatic rewriting of marketing content based on business needs
- Extracting product attributes from images and free text
- Detecting product variants
- Product categorization
- Automated onboarding of sellers’ products
- Merging product pages from multiple sources
- Predicting trending products
What’s in it for you:
- Build algorithms that visibly impact 500+ e-commerce/marketplace sites in 40 countries, including some with very high volumes (millions of products, customers, and orders per year)
- Work with cutting-edge techniques (multimodal models, LLM fine-tuning, etc.). Mirakl is one of the few French players with fine-tuned LLMs in large-scale production. Join us and keep pushing that pioneer spirit
- Real autonomy and ownership over your projects
Our stack and tools:
Python, Tensorflow, Pytorch, Hugging Face, Databricks, Spark, AWS (Amazon Redshift, s3, etc.), SQL, Airflow, Delta Lake. Spécifiques LLM : Autotrain, Unsloth, Galileo, LangChain, Anyscale.
Day to day, you will:
- Analyze and prepare data, prototype algorithms
- Put them into production with Data Engineers and dev teams
- Build dashboards to demonstrate algorithm performance and monitor production
- Present results at the weekly data science meeting and join team brainstorms
- Partner with other teams to refine use cases, user experience, and integration paths
You’ll love this job if:
- You have at least 4 years’ experience as a Data Scientist, with strong hands-on NLP and applied ML in industry
- You’ve deployed Machine Learning algorithms to production
- You know NLP and Computer Vision algorithms and state-of-the-art architectures (e.g., Transformers). Knowledge of the latest LLMs is a plus
- You’re fluent in Python and TensorFlow and/or PyTorch
- You have experience with Spark development
- You’re pragmatic, data-driven, and business-oriented
- You take full ownership of your topics, work autonomously, and are a great team player
- You bring a positive mindset: respect and kindness are core to your values
- You enjoy sharing your work through internal talks, conferences, or writing
Meet Arthur Delaitre, Data Science Manager for the team:
Wants to join us ? ⭐
- A 30-minute phone call with one of our Tech recruiters. We’ll discuss your background, expectations, and what Mirakl can offer you
- A 30-minute technical Zoom with someone from the Data Science team to dive into concrete aspects of your expertise and how it fits our projects
- A take-home assignment
- A 75-minute technical debrief and discussion with the Data Science team manager
- A final 1-hour Zoom with future Mirakl colleagues about our values and culture
We welcome collaborators with their diverse perspectives and experiences to power us forward. These often far exceed conventional job requirements and help us create a culture of continuous learning. If you’re ready to join a global leader powering digital transformation for 450+ of the world’s most innovative retailers and B2B organizations..
We may use Artificial Intelligence (AI) solutions to help streamline our hiring process, including screening applications, analyzing resumes, and assessing responses. While AI helps us work efficiently, all final hiring decisions are made by humans. For more information, visit our AI Guidelines for Candidates and Interviews.
Создайте идеальное резюме с помощью ИИ-агента

Навыки
- Python
- TensorFlow
- PyTorch
- Hugging Face
- Databricks
- Spark
- AWS
- SQL
- Airflow
- Delta Lake
- LangChain
- NLP
- Computer Vision
- Generative AI
- LLM
- Transformers
Возможные вопросы на собеседовании
Вакансия сфокусирована на каталогах, где данные часто зашумлены. Важно понять, как кандидат справляется с очисткой данных.
Как бы вы подошли к задаче дедупликации товаров от разных продавцов, если описания написаны на разных языках и имеют разную степень детализации?
Mirakl гордится использованием fine-tuned LLM в продакшене. Вопрос проверяет практический опыт оптимизации.
Какие техники оптимизации (например, Quantization, LoRA) вы использовали при деплое LLM для снижения задержек и стоимости инфраструктуры?
В стеке указан Spark, что критично для работы с миллионами товаров.
Опишите ваш опыт работы со Spark для обработки неструктурированных текстовых данных. С какими узкими местами в производительности вы сталкивались?
Роль предполагает тесное сотрудничество с продуктовыми командами.
Как вы оцениваете бизнес-эффективность (ROI) ваших моделей NLP после их внедрения в продакшен? Приведите пример метрик.
Проверка знаний современных архитектур, упомянутых в описании.
В чем преимущество использования мультимодальных моделей по сравнению с чисто текстовыми при извлечении атрибутов товара из каталога?
Похожие вакансии
Senior Data Scientist
Senior Data Scientist
Senior Data Scientist
Senior Data Scientist
Senior MLOps Engineer (ML Workflows Engineering)
Senior ML Engineer (JetBrains Research)
1000+ офферов получено
Устали искать работу? Мы найдём её за вас
Quick Offer улучшит ваше резюме, подберёт лучшие вакансии и откликнется за вас. Результат — в 3 раза больше приглашений на собеседования и никакой рутины!
- Страна
- Франция