Data Scientist / AI Engineer

NES Fircroft; Paris; Hybrid; EUR 90,000 - 130,000; Full time

Job summary

NES Fircroft in Paris is seeking a Data Scientist / AI Engineer for a 12-month contract. Hybrid work (3 days on site, 2 days remote) with fluent French and English requirements. Start date 2 November 2026. You will frame AI problems, explore data, and industrialize models with MLOps practices.

The role focuses on anomaly detection, forecasting, time-series handling, and GenAI / agentic systems within a Data+AI team for industrial use cases in oil and gas operations. CVs must be in English.

Qualifications

• Minimum 5 years in data science / AI engineering with production models.
• Strong Python ecosystem experience (pandas, NumPy, scikit-learn, PyTorch).
• Experience with end-to-end ML: data prep, evaluation, deployment and monitoring.

Responsibilities

• Frame use cases with business and product teams and define metrics.
• Explore data sources and build evaluation sets with experts.
• Develop, compare and industrialize models with MLOps in production.
• Design GenAI/agentic bricks and ensure traceability of results.

Skills

Python

ML

Time Series

MLOps

Cloud platforms

Tools

MLflow

Git/GitHub

Databricks

Docker

FastAPI

Job description

Data Scientist / AI Engineer – 12 Months Contract – Paris, France

Hybrid 3 days on site 2 days remote

Must speak fluent French with a good command of English

CV’s have to be in English

Start date 2nd November 2026

Required: - Minimum 5 years in data science / AI engineering, including at least 2 years on industrial projects with models deployed & maintained in production. Python & scientific ecosystem / ML (pandas, NumPy, scikit-learn, PyTorch). Modeling of industrial time series: anomaly detection, multivariate analysis, forecasting, management of noisy & irregular sensor data. Design of rigorous evaluations: business metrics, baselines, false positive/false negative arbitrage, reproducibility of benchmarks. End-to-end industrialization of ML solutions: MLOps (MLflow or equivalent), CI/CD with Git / GitHub, packaging, supervision. GenAI & agentic systems: RAG, structured outputs, tool calls, agent orchestration, LLM evaluation & traceability. Cloud & data platforms: Databricks, Azure &/or AWS; SQL; APIs (FastAPI); Docker.

Project or Product Context:

The service is part of the Data+AI team of the Digital Factory (AI for industry). The team has a portfolio of industrial AI use cases, particularly in the context of E&P's Digital Plant programs (e.g. condition-based maintenance, production optimization) & mobilizes its data scientists flexibly according to program priorities. The consultant will therefore be involved in several use cases during the service, & not on a single product, in conjunction with business experts (maintenance, reliability, production, processes), product teams & technology partners. Typical issues include (but are not limited to):
Anomaly detection & alert qualification on sensor data, reduction of false positives, prediction of process or production variables. Decision support based on equipment & maintenance history, identification of failure modes, rationalization of maintenance plans. Extracts structured information from reports, maintenance logs, & technical documents. Quantity estimation, monitoring automation, prioritization of opportunities or interventions. Tooled agents & AI assistants combining data, models & business knowledge, with traceable results.

Main missions: Frame use cases with the business & product teams: translate the need into a data/AI problem, define hypotheses, business metrics & success criteria. Explore & qualify data (time series, SAP maintenance data, documents, production data) & build evaluation sets with experts. Design, develop & compare models against solid business baselines (classical ML, statistical methods, foundation models for time series, LLM agents). Design GenAI / agentic bricks when relevant (historical case research, structured extraction, tooled agents), with traceable results. Industrialize: reproducible pipelines, packaging, APIs, deployment & supervision in production (MLOps / LLMOps), according to the team's standards. Ensure the life cycle of the models: performance monitoring, drift detection, retraining, documentation. Deliver results & limits to business users & management & contribute to go/no-go recommendations. Deliverables (defined by subject with the use case holder):
Scoping notes & evaluation protocols (metrics, baselines, datasets). Exploration/benchmark reports with recommendations. Versioned & tested code, templates, agents, or deployable API components. Documentation of production monitoring models & indicators.

Nice to have: Business knowledge: predictive/condition-based maintenance, reliability (failure modes, RCA), oil & gas production operations. Industrial data systems: PI / historians, Inmation, SAP PM, Cognite Data Fusion. Documentary/vision AI (P&ID, technical plans). Critical use of AI in development practices, without delegating understanding or technical responsibility.

With over 90 years' combined experience, NES Fircroft (NES) is proud to be the world's leading engineering staffing provider spanning the Oil & Gas, Power & Renewables, Chemicals, Construction & Infrastructure, Life Sciences, Mining andManufacturingsectors worldwide. With more than 80 offices in 45 countries, we are able to provide our clients with the engineering and technical expertise they need, wherever and whenever it is needed. We offer contractors far more than a traditional recruitment service, supporting with everything from securing visas and work permits, to providing market-leading benefits packages and accommodation, ensuring they are safely and compliantly able to support our clients.