Postdoctoral researcher in hybrid modelling of anaerobic digestion using ADM1 and machine learning

Postdoctoral researcher in hybrid modelling of anaerobic digestion using ADM1 and machine learning

Context

The shADow project proposes a digital shadow approach for anaerobic digesters, leveraging a hybrid model that combines ADM1 - the reference model for anaerobic digestion - with machine learning. This approach will improve biomethane production predictions through enhanced integration of physico-chemical and biological parameters while ensuring process interpretability and adaptability to real-time variations.

In the first phase, different model variants will be developed and compared using available physico-chemical and biological data from anaerobic digesters. In the second phase, semi-continuous digesters and their digital shadow will be implemented to evaluate predictive performance. A key focus will be assessing the impact of microbial data integration on prediction accuracy under inhibited conditions, where ADM1-based predictions are typically less
reliable than under standard conditions. By optimizing anaerobic digestion through digital modeling, shADow takes a step towards the implementation of digital twins.

Work environment, missions and activities

The postdoctoral researcher will be hosted at the Toulouse Biotechnology Institute (TBI) a public research laboratory on the INSA Toulouse campus, jointly supported by INSA Toulouse, CNRS and INRAE. The position will be based within TBI’s Mathematics Cell, which focuses on mathematical modelling of biochemical and microbial systems and on AI/ML methods. The postdoctoral researcher will collaborate closely with experts from INRAE’s national laboratories, including the LBE lab in Narbonne (dynamic modelling) and PROSE in Antony (experimentalists and data to use in this project). The position will involve research stays at both laboratories in
order to strengthen collaboration and integrate experimental and modeling approaches.

Missions & activities

• Develop and refine ADM1-based models using laboratory and industrial datasets, with a focus on robust parameter estimation, sensitivity analysis, and uncertainty quantification.
• Design, test and compare hybrid modeling strategies that combine mechanistic ADM1 models with data-driven machine learning approaches.
• Apply machine learning methods (e.g., regression, sequence models, surrogate models) to improve prediction, state estimation, and scenario analysis.
• Implement and promote MLOps best practices, including reproducible workflows, version control, automated testing, and model deployment pipelines.
• Collaborate closely with experimentalists and process engineers to integrate data and provide actionable insights.
• Disseminate research results through publications, conferences, and project deliverables.

Experience

• PhD in chemical engineering, environmental engineering, applied mathematics or a related field.
• Solid background in process modelling and simulation, ideally with experience in ADM1 or similar mechanistic models.
• Proven experience with machine learning methods applied to time series, system identification, or process modelling.
• Strong programming skills in Python, MATLAB, or similar environments.
• Track record of scientific publications in relevant fields.

What We Offer

• Mentorship and training: receive guidance from leading experts in the field and access to a comprehensive training program.
• Collaborative network: engage with a dynamic team of researchers and industry partners, both nationally and internationally.
• Career development: opportunities for professional development, including workshops, seminars, and networking events.
• Work atmosphere: You will be part of a team that takes their work very seriously. We pride ourselves on fostering a collaborative and supportive atmosphere where researchers and interns work together in a dynamic and friendly environment to drive innovative solutions for environmental sustainability.
• Great working conditions: 30 days of holidays per year (+ 15 if time recovery is affected); support for parenting; access to inexpensive collective restaurants; flexible schedule; access to on-site sports facilities; sunny weather.
• Salary (gross): 2550 - 2 840 €/month (including full social and health benefits).

Personal skills

• Strong motivation to work in an interdisciplinary and collaborative research environment.
• Ability to work independently and manage research tasks effectively.
• Excellent problem-solving and analytical thinking skills.
• Good communication skills in English (spoken and written).
• Open-minded, creative, and eager to contribute to both fundamental and applied research.
• Academic Background: PhD degree in Control, Process Engineering, Computer Science, Mathematics (or related field), or an equivalent engineering degree
• Skills (recommended): experience in bioprocess modelling; scientific communication and writing skills.
• Language: fluent use of English; knowledge of French is an advantage but not required.

Location: Toulouse Biotechnology Institute (TBI). 135 Av. de Rangueil, 31400 Toulouse.
Contract: 18 months contract (CDD).
Start date: as soon as possible

Contact:

Interested candidates should submit the following documents:

• A detailed CV, including academic achievements and research experience.
• A cover letter outlining your motivation and suitability for the project.

To:
DavidCamilo.CorralesMunoz@inrae.fr
Elie.Le-Quemener@inrae.fr
gabriel.capson-tojo@inrae.fr
ariane.bize@inrae.fr
jean-philippe.steyer@inrae.fr
olivier.chapleur@inrae.fr