Senior Data Scientist
Pubblicato il 14-09-2026 - Welyk in Torino
Digital Pills is a consulting company specialized exclusively in data, covering the full data value chain: data strategy, data engineering, analytics engineering, BI and data science.
The company works across two main areas, reducing customer acquisition costs and increasing customer lifetime value, for clients ranging from large international brands (including HelloFresh, Last Minute, Dolce & Gabbana, Arcaplanet) to non-profit organizations such as AIRC and Save the Children, to which the company donates 1% of its annual revenue.
Despite being a company of around 30 people, Digital Pills has recently started an international expansion with a hub in Berlin and is in the middle of a strategic transition: from bespoke consulting to a scalable model built on proprietary products and accelerators, developed starting from projects with its most strategic clients.
What you'll do
- Design, develop and validate solutions across classical machine learning, applied AI (agents, copilots, LLM-based services) and advanced statistics, bringing them to production with the right level of rigor
- Choose the right method for each context among experimentation, A/B testing, backtesting and causal inference, proving the business impact of what you build
- Own 2-3 client projects end-to-end in parallel, including relationships with senior client stakeholders
- Identify recurring patterns across projects and abstract them into reusable frameworks and accelerators, in collaboration with the CTO
- Mentor junior data scientists and contribute to the team's training roadmap
- Support Client Managers in pre-sales scoping, translating business requirements into technical architectures and realistic estimates
- Bring solutions to production in close collaboration with Data Engineering and Analytics Engineering colleagues
- Translate technical work into clear narratives for non-technical stakeholders
Technical requirements
- Strong Python (pandas, NumPy), ML/DL frameworks (scikit-learn, PyTorch