Principal AI Product Manager – AI Catalog
Location: Paris, Bordeaux, or Remote (France)
At Mirakl, we are embedding state-of-the-art foundation models into our SaaS platform to radically transform product catalog operations at global scale. We’re hiring a Principal AI Product Manager to lead the technical product strategy and execution for our AI Catalog Product (Catalog Transformer).
In this senior IC role, you will work hands-on at the intersection of LLM architecture, data pipelines, and user-facing SaaS products — translating research innovations into production-grade, scalable AI features. You will define how foundation models (OpenAI, Mistral, Claude, Llama) are fine-tuned, orchestrated (e.g., via RAG pipelines), evaluated, and deployed into critical catalog data transformation tasks.
You’ll collaborate closely with AI Research, Data Science, Engineering, and SaaS Product teams (Catalog & Connect), and you’ll lead from the front in ensuring AI outputs meet industrial-grade standards of factuality, consistency, latency, and robustness.
What you will do:
Own end-to-end AI product strategy and execution for Catalog transformer (attribute extraction, translation, enrichment, and media improvement & moderation) across Mirakl Connect and Mirakl Catalog products
Drive cross-functional collaboration with AI researchers / Datascientist, PM, designers, and engineers to deliver impactful features.
Own the integration and evaluation of GenAI models, ensuring quality, reliability, and performance.
Drive experimentation with GenAI tooling and scalable deployments.
Collaborate on AI infra choices (model serving, retrieval stores, quality monitoring).
Build internal tooling and evaluation pathways to enable Sales and CS teams to showcase GenAI capabilities.
Who you are:
7+ years of SaaS product management, with deep technical ownership over AI/ML features.
MSc/PhD in Computer Science, Machine Learning, or equivalent hands-on expertise.
Strong technical fluency in GenAI ecosystems: LLMs, fine-tuning, RAG systems, embeddings, structured output modeling.
Hands-on experience working with AI/ML teams on model lifecycle: evaluation, A/B testing, fine-tuning, and post-deployment monitoring.
Analytical and experimental mindset: you know how to validate an LLM product beyond demos.
Familiarity with computer vision pipelines, data labeling ops, or human-in-the-loop AI systems.
Fluent in French and English.
Practical experience in catalog data management, search/NLP, or eCommerce AI systems is a plus
(Gen)AI at the Heart of Mirakl's ProductAnne-Claire Baschet, Chief Data & AI Officer at Mirakl, discusses the journey from inception to deployment of the Catalog Transformer, highlighting the integration of GenAI into Mirakl's core products.
Watch here: YouTube
Building and Scaling an AI Product in a SaaS SolutionFanny Guélin and Arthur Delaitre from Mirakl present at Apidays Paris 2024, discussing the development and scaling of AI products within a SaaS framework, with a focus on the Catalog Transformer.
View the presentation: Speaker Deck