Scientific & technological research project
Artificial intelligence for the modelling and optimisation of dough-based bakery product recipes
SMART BAKE AI connects organised experimental production data with algorithmic machine learning methods, so that the development of a recipe no longer rests solely on experience and on repeated trial production runs.
| Beneficiary | Charoula Ioanna Tsitsipa — sole proprietorship, Larissa |
|---|---|
| Implementation period | 01.01.2025 — 31.01.2026 (13 months) |
| Scientific lead | Charoula Ioanna Tsitsipa |
| Framework | Self-funded project, Article 22A of Law 4172/2013 |
13
months of research work
4
work packages (WP1–WP4)
3
online deliverables
1
experimental artificial intelligence application
The research problem
Why a recipe does not behave predictably
The initial idea for the project arose from the established difficulty of predicting the behaviour of a dough-based bakery product during the production process.
A multitude of interdependent parameters
The type and proportion of raw materials, the dosage of the ingredients, the fats used, the moisture, the baking temperatures and times, the production and storage conditions.
Non-linear interactions
Small changes in the parameters may bring about significant differences in texture, in volume, in colour, in thermal behaviour, in the appearance of spoilage and in shelf life.
Limited generalisability
The conclusions drawn from one recipe do not necessarily apply reliably to different categories of dough-based bakery products.
The scientific and technological uncertainty. At the start of the project it was not known in advance which input features, which form of pre-processing of the data and which algorithmic approaches would provide reliable results — nor whether these relationships could be captured at all with sufficient accuracy through artificial intelligence models.
Research methodology
From experimental production to the algorithmic model
The methodology was iterative in character: the results of the initial models were not treated as definitive, but were evaluated and, where necessary, rejected.
Structure of the project
Four functionally interconnected work packages
WP1
Construction of experimental test specimens and data collection
The experimental basis of the project. Systematic production of test specimens and creation of an organised set of real experimental data.
01.01.2025 – 31.12.2025
WP2
Scientific investigation and initial technological design
Evaluation of the data, determination of variables and uncertainties, design of the model architecture.
19.11.2025 – 30.11.2025 · D1
WP3
Development and evaluation of an experimental software version
Training, retraining and comparative evaluation of models; integration into an experimental web application.
01.12.2025 – 31.12.2025 · D2
WP4
Final evaluation, documentation and publication
Synthesis and assessment of the results, recording of limitations and of the new scientific and technological knowledge.
01.01.2026 – 31.01.2026 · D3
Deliverables
Published research results
The dissemination of the results was carried out through a purpose-built online research environment, organised into three distinct sections.
D1
Online documentation of the research methodology and of the initial experimental results
Methodology, evaluation of data quality, input variables, initial uncertainties and design of the algorithmic approach.
Delivered 30.11.2025 smartbakeai.com/research/documentation
D2
Experimental artificial intelligence software version for the modelling and optimisation of recipes
A functional application for entering selected parameters and producing experimental predictions, with documentation of inputs, outputs, metrics and limitations.
Delivered 31.12.2025 smartbakeai.com/research/software
D3
Online final evaluation and results report of the scientific and technological research
Synthesis of the experimental results, accuracy and generalisation metrics, unsuccessful approaches, technical limitations and final conclusions.
Delivered 31.01.2026 smartbakeai.com/research/results
Research team
Who implemented the project
| Member | Capacity | Contribution to the project |
|---|---|---|
| Charoula Ioanna Tsitsipa | Owner · head of the research and development department | Scientific and administrative responsibility; definition of research objectives and priorities, coordination of the experimental process, supervision of test specimen production, review and acceptance of the deliverables. |
| Eirini Tsitsipa | Administrative staff · MSc, Division of Psychiatry, University College London | Design and execution of experimental trials, monitoring of composition and production parameters, collection and organisation of data, evaluation of quality characteristics. |
| Andriana Oikonomou | Technical staff | Preparation of raw materials, daily preparation of test specimens, execution of trials and recording of quantities and production conditions. |
| ALUMICAD Single-Member P.C. Scientific lead: Panagiotis Karakitsios | Subcontractor · deep technology startup Dipl. Civil Engineer NTUA, interdepartmental MSc “Structural Design and Analysis of Structures”, specialising in computational mechanics and artificial intelligence | Design of the research and computational methodology, data pre-processing, architecture of the algorithmic models, development and retraining, comparative evaluation, development of the experimental application and technical documentation. |
The collaboration with ALUMICAD Single-Member P.C. was based on a contract for the provision of scientific and technological research services of 19 November 2025.
Interested in the methodology?
Larissa · 6955352982 · xaroulatsitsipa@gmail.com