CIVINTEL
ELEN
Contact

Discipline 01

AI & Data Science

Machine learning that works on the factory floor, not only in the lab.

We work with owners, technical directors, R&D, IT and production teams to turn operational questions into measurable AI pilots, and successful pilots into tools people use every day. Our experience spans aluminium production, fragrances and chemicals, cosmetics, food, agriculture, public health and semiconductors.

Who we work with

  • Metals & aluminium
  • Chemicals & fragrances
  • Cosmetics & consumer goods
  • Food industry
  • Agriculture
  • Public health
  • Energy-intensive processes
  • Research organisations

Services

What we take on

01

AI strategy & use-case discovery

Workshops with management, R&D and IT to find where AI is worth applying, what data exists and what a first pilot would need.

Typical deliverables

  • Use-case map ranked by value and feasibility
  • Data readiness assessment
  • Pilot brief with success criteria and cost estimate
What we need from you
  • A sponsor from management and access to R&D, IT and production
  • An overview of processes and available data (no raw data needed at this stage)
Typical duration
2–4 weeks
Scope & limits
An assessment and roadmap, not an implementation. Data is reviewed at summary level.
02

Computer vision & quality inspection

Automated visual inspection of products, surfaces and processes with industrial cameras and deep learning.

Typical deliverables

  • Defect detection and classification models
  • Inspection interfaces for operators
  • Event history and quality reports
What we need from you
  • Sample images, or access to the line to capture them
  • Examples of acceptable and defective products and the acceptance rules
  • Time from a quality expert to review labels
Typical duration
Pilot: 8–16 weeks
Scope & limits
Accuracy depends on lighting, camera position and labelled examples. New defect types require retraining.
03

Predictive models & optimisation

Models that forecast consumption, quality or risk, and optimise recipes, formulations or operating parameters.

Typical deliverables

  • Validated model with error analysis
  • Explainable predictions and recommendations
  • Digital-twin scenarios for “what-if” questions
What we need from you
  • Historical data linked to the outcome you want to predict or optimise
  • A domain expert to validate results
Typical duration
Pilot: 8–12 weeks
Scope & limits
Models are valid within the range of the data they were trained on. Recommendations are reviewed by people before use.
04

Data engineering & analytics

We connect ERP, SAP, sensors, APIs and spreadsheets into clean, linked datasets and turn them into indicators your team trusts.

Typical deliverables

  • ETL pipelines and data models
  • KPI definitions and benchmarks
  • Data-mining and statistical analyses
What we need from you
  • Read access to data sources (ERP/SAP exports, databases, APIs, sensors)
  • A contact person in IT
Typical duration
4–10 weeks
Scope & limits
We connect to your ERP or MES; we do not replace them.
05

Dashboards & custom software

Full-stack web applications and interactive dashboards built with React, Node.js, Grafana and modern charting libraries.

Typical deliverables

  • Web dashboards with filters and KPIs
  • APIs and integrations
  • Deployment on your infrastructure
What we need from you
  • An agreed list of KPIs and users
  • A server or cloud account for deployment, or we propose one
Typical duration
4–12 weeks
Scope & limits
Hosting and third-party licences are billed separately where needed.
06

Training & R&D support

Industrial AI seminars for technical teams, and technical support for research and funded innovation projects.

Typical deliverables

  • Tailored seminars with examples from your sector
  • Technical sections for funding proposals
  • Scientific review of AI approaches
What we need from you
  • Audience profile and topics of interest
  • Optionally, anonymised examples from your company
Typical duration
From one day to several weeks
Scope & limits
For funding proposals we prepare the technical sections; the applicant remains responsible for the submission.

Have a question your data or your building could answer?

A first conversation is free and without obligation. Tell us what you want to achieve and we will tell you honestly whether and how we can help.