Independent expertise · Food & biotech
From data
to insight.
Scientific data & AI engineering for food and biotechnology.
I’m George Pampoukis, PhD. I turn complex scientific data into usable tools and clearer decisions for industry and research teams.
Zurich, Switzerland · Industry & university collaborations
Expertise
Built around
your question.
Scientific data workflows, analysis, and tools—scoped to the decisions you need to make.
- Connect and reuse scientific data. Data management and engineering: clean, combine, and validate experiment files and metadata through repeatable workflows.
- Compare experiments and guide optimization. Statistics, interactive plots, and predictive models to compare batches or conditions, with explicit validation and uncertainty.
- Put literature data to work. AI-assisted extraction into reviewed, source-linked datasets for modeling and scenario comparisons.
- Quantify microscopy images. Cell counting, time-lapse phenotyping, and growth-state classification, with human review of ambiguous results.
Interactive workspace
Find the balance.
Not just the maximum.
Tables to editable data. Data to predictions. Competing outcomes to a better next experiment.
Source tables Three documents, different column layouts
| Source | °C | pH | Min | Yield / activity | Functionality / stability |
|---|---|---|---|---|---|
| Food / protein extraction yield & functionality | |||||
| A.1 | 25 | 5 | 15 | 22 | 80 |
| A.2 | 25 | 7.5 | 40 | 36 | 62 |
| A.3 | 35 | 4.5 | 30 | 30 | 69 |
| A.4 | 35 | 6.5 | 15 | 42 | 88 |
| B.1 | 40 | 8 | 50 | 70 | 36 |
| B.2 | 45 | 5.5 | 40 | 58 | 53 |
| B.3 | 45 | 7 | 25 | 63 | 71 |
| B.4 | 50 | 6 | 30 | 67 | 61 |
| C.1 | 55 | 8.5 | 20 | 75 | 54 |
| C.2 | 60 | 6.5 | 50 | 82 | 24 |
| C.3 | 65 | 5 | 55 | 61 | 18 |
| C.4 | 70 | 7.5 | 10 | 59 | 33 |
| Biotech / enzyme activity & stability | |||||
| A.1 | 25 | 5 | 15 | 36 | 86 |
| A.2 | 25 | 7.5 | 40 | 32 | 62 |
| A.3 | 35 | 4.5 | 30 | 34 | 75 |
| A.4 | 35 | 6.5 | 15 | 67 | 89 |
| B.1 | 40 | 8 | 50 | 51 | 38 |
| B.2 | 45 | 5.5 | 40 | 65 | 58 |
| B.3 | 45 | 7 | 25 | 86 | 69 |
| B.4 | 50 | 6 | 30 | 81 | 57 |
| C.1 | 55 | 8.5 | 20 | 39 | 33 |
| C.2 | 60 | 6.5 | 50 | 54 | 19 |
| C.3 | 65 | 5 | 55 | 21 | 22 |
| C.4 | 70 | 7.5 | 10 | 34 | 29 |
Editable data
Edit values. Both responses and contours refit live.
12 source-linked rows / 2 fits
Predictions & trade-offs
No contour at 10-index intervals crosses this slice.
Two quadratic fits; weighted response sum. Grid search: 1 °C, 0.1 pH, 5 min. Indices, not measured percentages or validated process recommendations.
Cell morphology: size and shape
Count by size and shape
Explore microbial morphology in a synthetic field. Adjust the size filter and the round/elongated split.
Cell 3: not flagged.
Size = equivalent-circle diameter. Shape factor is circularity: 1 = round, not a species label.
Drawn cells only; no image analysis.
Independent expertise
Science, connected
to usable tools.
My doctoral work connected data-driven methods with food research. I now work across scientific data engineering, machine learning, and AI-assisted tools for industry and university teams.
Food and biotechnology are my core focus. I also welcome related scientific data-management, data-engineering, and optimization projects.
We define the question, test the work together, and agree on documentation and handover.
Selected research
Doctoral thesis · Wageningen University · 2026
From data to insight (opens in a new tab)Assessing the decontamination efficacy of mild food processing technologies using data-driven approaches.
-
A metadata schema to standardize non-thermal plasma decontamination parameters in food-related applications
Scientific metadata and reusable research data.
-
Ranking factors affecting the decontamination efficacy of non-thermal plasma: The approach of dissipated power per plasma volume through machine learning modeling
Machine learning for a food-processing question.
-
The decontamination efficacy of pulsed electric field in food: Establishing global kinetic parameters
Quantitative modeling in food microbiology.
Tell me about
your data.
Tell me what data you have, what gets in the way, and the decision or workflow you want to improve.
Connect on LinkedIn (opens in a new tab)Please avoid confidential or sensitive information in an initial inquiry.