Data Scientist

Le contenu du poste est libellé en anglais car il nécessite de nombreuses interactions avec nos filiales à l’international, l’anglais étant la langue de travail.

• Location: Lyon, M&S Accelerator
• Work Arrangement: Hybrid (Three days from office and two days from home)

Job title

Data Scientist

About The Job

As a Data Scientist within our Digital Manufacturing & Supply (M&S) Accelerator, you'll take ownership of turning real-time signals from our factories and labs into decisions that matter — working at the intersection of drug development, advanced manufacturing, and cutting‑edge AI. Ready to get started?

Join Sanofi's Manufacturing & Supply team right in the heart of Lyon, in a vibrant startup-style environment in the city centre. You'll support 12 blockbuster launch teams and 40 global factories across R&D and M&S, helping accelerate the launch of new medicines, build digitalised "Lighthouse" factories, and drive manufacturing performance improvements — all while enjoying an easy commute and a great work-life balance in one of France's most dynamic cities.

Main Responsibilities

• Partner with business teams to understand requirements and translate them into technical solutions, working with a good degree of autonomy
• Own the ingestion, cleaning, and harmonisation of IoT and sensor data from a mixed estate — some fully connected and streaming, plenty of it legacy, manual, or messy. A meaningful part of the role is data cleaning and wrangling: it's the foundation everything is built on, and doing it well is genuinely high-impact here
• Design and deliver time-series models — forecasting, anomaly detection, and monitoring — on streaming sensor data to drive manufacturing performance and quality
• Design, build, and support real-time and streaming data pipelines together with engineers and MLOps, and help bring legacy sites up to a usable state
• Write highly optimised, production-ready code, taking models from experiment through to deployment
• Scope, define, and deliver AI-based data products end to end, using data analysis, visualisation, and storytelling
• Be accountable for the robustness of the AI solutions you deliver, applying responsible AI practices (monitoring, explainability, trustworthiness, ethics)
• Contribute to — and help raise the bar within — the global data science community of practice
• Support junior data scientists and analysts with informal guidance and code review

You'll also work across the wider AI portfolio — including LLMs, Retrieval-Augmented Generation (RAG), and agentic AI workflows — as our products increasingly blend sensor intelligence with GenAI. Strong here is a plus, but IoT and time-series is the core of this role.

About You

You're a capable data scientist with a few years under your belt, genuinely excited by physical-world data — sensors, signals, and the machines that produce them. You can take a problem from vague business need to deployed model with limited hand-holding, and you've got a keen eye for improvement opportunities.

Experience

• Hands‑on track record working with IoT data, sensor streams, or time-series (this is what we care about most)
• Demonstrated experience applying machine learning to real problems, ideally involving signals or temporal data
• Experience developing deployable code and taking models to production in an agile environment
• You're not put off by messy, real-world data — a big part of this role is turning inconsistent legacy data into something usable, and doing it well is where a lot of the value sits
• Knowledge of the pharmaceutical or chemical domain is a huge advantage — understanding of manufacturing processes, GxP, or lab environments will set you apart

Soft Skills

• Excellent written and verbal communication skills
• Experience working with multiple teams to drive alignment and results
• Service-oriented, flexible, positive team player
• Self‑motivated, takes initiative
• Problem solving & critical thinking

Technical Skills

• Strong Python, and comfortable with at least one database system — time-series databases (e.g. InfluxDB, TimescaleDB) especially valued
• Solid command of time-series methods — forecasting, anomaly detection, signal processing — on streaming data
• Working knowledge of streaming / real-time data technologies (e.g. Kafka, Spark Streaming, MQTT or similar)
• Strong ML foundations: supervised/unsupervised learning, deep learning, with a solid grasp of the underlying theory
• Comfortable working in cloud and high-performance computing environments (AWS, GCP, Databricks, Apache Spark)
• Good software engineering practices — version control, testing, CI/CD, orchestration
• Experience with data visualisation tools (Powe