Data Scientist
Summary
Hybrid Lisbon data scientist building predictive price/demand forecasting and market-signal models for a major utilities client, working with large time-series datasets and using generative AI tools (LLMs, RAG, agentic workflows) alongside the Python ML stack (pandas, scikit-learn, XGBoost/LightGBM, PyTorch/TensorFlow).
Data Scientist – Predictive Modeling & Generative AI
Location: Lisbon
Work Model: Hybrid (1-2 days per week)
About the Role
We are building a team to develop predictive algorithmic models for a major utilities company. This is a high-impact, applied data science role where you'll work at the intersection of machine learning, generative AI, and quantitative modeling to help anticipate market movements and price dynamics.
What You'll Do
- Design, build, and validate predictive models for price forecasting, demand/supply dynamics, and market signal generation
- Use generative AI tools (LLM-assisted coding, RAG pipelines, prompt engineering, agentic workflows) to speed up research, feature engineering, and model iteration
- Work with large, messy, time-series-heavy datasets specific to our client's market
- Translate trading team needs into modeling problems, and translate model outputs back into insights that can be acted on
- Continuously test, monitor, and refine models in a live, fast-moving market environment
- Collaborate with client engineers to embed models into production workflows
- Document methodology and findings clearly for both technical and non-technical stakeholders
What We're Looking For
Experience & Technical Skills
- 3–5 years of experience as a Data Scientist, ML Engineer, or similar applied role
- Strong hands-on experience with Python and standard ML/data stacks (pandas, scikit-learn, XGBoost/LightGBM, PyTorch or TensorFlow)
- Practical experience using generative AI / LLM tools in a real workflow (e.g., Copilot-style coding assistants, LangChain/LlamaIndex, RAG, prompt engineering, agentic pipelines) — this is a hands-on requirement, not just familiarity
- Solid grounding in predictive modeling and time-series forecasting techniques
- Comfortable working with SQL and large, real-world (i.e. imperfect) datasets
- Experience deploying or maintaining models in a production or near-production setting is a plus
Domain Knowledge
- Any exposure to commodities, trading, financial markets, or energy systems is a strong plus, not a requirement
- A genuine will to learn a complex, fast-moving business model is essential
Soft Skills & Working Style
- High degree of autonomy — comfortable owning problems end-to-end with limited hand-holding
- Strong intellectual curiosity and drive to grow technically and domain-wise
- Clear communicator, able to work closely with traders and non-technical stakeholders
- Comfortable in a hybrid, project-driven environment with evolving priorities