Forward Deployed Engineer - Machine Learning (F/H) - CDI

Collective.work; Paris; Hybride; EUR 70 000 - 110 000; Plein temps

Avantages offerts par ce poste

Health insurance

Meal vouchers

Equity plan

Résumé du poste

Collective.work is hiring in Paris for a permanent ML Engineer focused on productionizing recommendation models and real-time services. You will own end-to-end ML lifecycle, from feature engineering to deployment and monitoring, while designing offline evaluations and scalable pipelines.

With 3+ years of production ML experience, you will collaborate with data engineers and product teams to optimize model performance, drift handling, and experiment design in a fast-growing marketplace.

Qualifications

• 3+ years of deploying and operating ML models in production.
• Strong Python skills and software engineering practices.
• Experience with PyTorch or TensorFlow and ML pipelines.

Responsabilités

• Build and maintain recommendation and ranking models for core product surfaces.
• Own the ML lifecycle from feature prep to deployment and monitoring.
• Design offline evaluation methods reflecting production behavior.

Connaissances

Python

ML engineers

Git

CI/CD

Experimentation

Outils

PyTorch

TensorFlow

CI/CD pipelines

Description du poste

Take recommendation models from notebook to production, serving millions of users in real time.

Paris, France · Permanent · Hybrid working policy to be confirmed

What you'd actually work on

• Building and maintaining recommendation and ranking models used across core product surfaces
• Owning the ML lifecycle from feature preparation and training to deployment and production monitoring
• Designing offline evaluation methods that reflect the behaviour expected in production
• Building training and retraining pipelines that can run reliably without manual intervention
• Deploying models through scalable, low-latency production services
• Designing A/B tests and defining the metrics used to assess model performance and business impact
• Working with data engineers to make production features reliable, consistent, and available at the required frequency
• Monitoring model performance, feature quality, drift, latency, and prediction distributions
• Investigating differences between offline results and production behaviour
• Improving model deployment, versioning, rollback, and reproducibility
• Partnering with product teams to translate business objectives into measurable optimisation problems
• Contributing to code reviews, automated testing, CI/CD, and ML engineering standards

Where it gets technically interesting

• Serving real-time features and predictions within strict latency requirements
• Maintaining consistency between offline training data and online production features
• Automating retraining pipelines while keeping model versions, datasets, and experiments reproducible
• Detecting model or feature drift before it has a significant impact on users
• Designing A/B tests with appropriate metrics, sample sizes, and evaluation periods
• Balancing model complexity and prediction quality against latency and infrastructure costs
• Handling cold-start problems for new users, products, or categories
• Rolling out new models safely, with clear monitoring and rollback mechanisms

What we're looking for

• 3+ years of experience deploying and operating machine learning models in production
• Strong Python skills and good software engineering practices
• Practical experience with PyTorch, TensorFlow, or an equivalent ML framework
• Experience building training, evaluation, and inference pipelines
• Knowledge of recommendation, ranking, personalisation, or similar machine learning systems
• Experience with model deployment, monitoring, versioning, and automated retraining
• Understanding of feature engineering and the differences between offline and online feature processing
• Experience designing or analysing controlled experiments and A/B tests
• Ability to investigate performance issues across models, data pipelines, and production infrastructure
• Experience using Git, code reviews, automated testing, and CI/CD
• Confidence working with data engineering, platform, and product teams

You do not need to have worked with every tool in the stack, but you should have experience taking models beyond experimentation and operating them as production systems.

The company

A fast-growing digital marketplace with millions of monthly active users, several hundred employees, and a strong presence across Europe.

Machine learning is used across key product areas, including recommendation, ranking, and personalisation. The team is developing the infrastructure and engineering practices required to operate these models reliably at scale.

• Health insurance, meal vouchers, and an equity plan.

Languages: Native or bilingual French and professional English.