AI Lead — PurePilot STEP

Main role

As the technical lead for the PurePilot STEP algorithmic component, your goal is to design, industrialize, and ensure the reliability of the optimization and control models deployed across hundreds of plants. You will bridge the gap between product requirements, the physical reality of the processes, and the state of the art in AI and advanced control.

Responsibilities 

1. Translate the product roadmap into algorithmic projects

• Work with the Product Owner to transform each product objective into a modeling problem, defining its data, constraints, and success criteria.
• Estimate, plan, and prioritize these projects.

2. Build the models

• Develop predictive models and data-driven digital twins for STEP processes: nitrification/denitrification, phosphorus treatment, and new wastewater treatment plant use cases.
• Automate model construction at scale: automatic site-specific selection and calibration (AutoML), and knowledge transfer between plants (meta-learning).
• Combine data with domain expertise where appropriate (hybrid data/physics approaches).
• Design control strategies: model predictive control (MPC), constrained optimization, and stochastic optimization.

3. Test before deploying

• Define metrics and validation protocols: historical replay, simulation, and pilot sites.
• Verify robustness: regime changes, missing or erroneous data, and rainfall events.

4. Deploy and maintain in production

• Industrialize models: ensure code is tested, versioned, and deployable across the entire fleet.
• Monitor performance in production and address drift; serve as the technical point of contact for algorithmic incidents reported by support.

5. Advance the team

• Share best practices, collaborate with other AI engineers, and mentor interns.
• Stay up to date with scientific research and propose R&D initiatives.

Education and experience:

• Master’s degree (engineering school or M2) or PhD in applied mathematics, control theory, computer science, or process engineering.
• At least three years of practical experience combining advanced dynamic system control and machine learning, with at least one proven project in a real industrial production environment.

Technical skills:

• Dynamic system identification, time-series modeling, and machine learning (scikit-learn, PyTorch, TensorFlow, or JAX)
• Probabilistic modeling: probabilistic forecasting, uncertainty quantification
• Optimal control: model predictive control, constrained optimization, stochastic optimization
• Advanced Python and software best practices: testing, code review, version control, continuous integration
• Ability to translate a concrete industrial problem into a modeling problem
• Technical English

Personal qualities:

• Industrial pragmatism: focus on tangible gains and on-site stability rather than algorithmic complexity alone.
• Communication and collaboration: ability to explain concepts clearly to non-specialists (product teams, automation engineers, field operators).
• Strong design autonomy and the ability to lead applied R&D projects from end to end.

Preferred qualifications (not mandatory):

• Knowledge of wastewater treatment or biological processes.
• Experience with MLOps architectures, containers (Docker), and edge/cloud deployment.
• Experience in transfer learning or meta-learning on heterogeneous fleets.

• Salary package: €35,000 to €45,000 range depending on experience.
• Work organization: flexible hours, hybrid remote work policy (three days on-site per week to encourage technical collaboration).
• Benefits: regular seminars, electric car-sharing fleet for business and personal travel, holiday bonus, and sustainable mobility allowance.