Data Scientist

CELSA es el primer productor europeo de acero circular de bajas emisiones y constituye la cadena de suministro circular más grande de Europa. Recicla chatarra férrica para producir acero en hornos de arco eléctrico, usando la tecnología más sostenible y la más eficiente energéticamente.

About CELSA

We’re Europe’s leading producer of circular, low-emission steel and its largest circular supply chains. We recycle ferrous scrap into new steel through electric arc furnaces, one of the most energy-efficient steelmaking routes available. Founded as a family business near Barcelona in 1967, the Group today brings together ~5,000 people and produces 5.7 million tons of steel annually, made with ~95% recycled content and designed to be 100% recyclable

Our purpose extends beyond steel: making the cleanest way to produce it also the smartest – and data, analytics and artificial intelligence are central to that ambition.

We are now embedding AI into core steelmaking operations, moving beyond isolated experiments toward models integrated into how steel is produced, traded and optimized. This role is suited to a hands-on builder who wants to see their work deployed in live operating environments, supporting furnace operators, metallurgists and traders

The Data Scientist can take models from concept to production in close collaboration with experts who understand steel’s physics, operations and economics. Working alongside our team, this role can help shape how the company makes decisions with data, building models that create operational and commercial impact.

What you’ll do

• Design, train and validate ML models for high-impact manufacturing, scrap, commercial and procurement use-cases

• Take models from prototype to production – feature pipelines, retraining, monitoring – deployed across cloud, edge and control-room

• Make models explainable and trusted: clear drivers, confidence levels, built for real adoption

• Partner with process experts and use-case owners so every model fits how the process actually runs

• Track impact against clean before-and-after baselines in monthly value reviews tied to P&L and cash

• Grow our reusable AI assets and ML environment, so each model makes the next one faster

What we’re looking for – must-haves

• A quantitative degree (Computer Science, Mathematics, Statistics, Physics, Engineering or similar), with experience to match: PhD + ~4 years, MSc + ~6 years, or BSc + ~8 years building and deploying ML models

• Strong Python and its ML ecosystem (scikit-learn, XGBoost, pandas), plus solid SQL and large-data handling

• Experienced in applying GenAI and LLM technologies to design, prototype, and deploy data-driven solutions

• Hands-on across the model lifecycle – experiment tracking, MLOps, retraining, monitoring

• Analytical rigor paired with business judgment – you care about the value a model creates, not just its accuracy

• Proactive and structured, comfortable explaining technical ideas to non-technical people

• High level of English

Bonus points

• Experience with time-series forecasting, optimization, or industrial / sensor data

• Databricks & Azure, or similar cloud ML platforms

• Exposure to deploying models in real-time, edge, or control-room settings

Not every requirement needs to be met in full. The criteria above reflect an ideal profile, while career paths often develop in less linear ways and strong candidates may not align with every item on a list. If this role is exciting and most of the profile feels aligned, we would value hearing from interested candidates rather than seeing them self-select out. We would welcome hearing what you have built.