Context
Within an Innovation team, I developed prototypes to collect and structure data from public procurement platforms and a CRM.
- CRM data
- Unstructured content
- Data collected from the web
- Data structured as JSON
Problem and challenge
The process needed to combine automated collection, content transformation and semantic comparison within an extensible architecture.
Objectives
- Automate data collection.
- Structure the information collected.
- Transform unstructured content into JSON.
- Compare the collected data semantically.
Contribution
I automated document retrieval with Python and Playwright, structured content into JSON and developed a semantic comparison system using LLMs.
- Automating collection with Python and Playwright.
- Structuring data from heterogeneous sources.
- Transforming content into JSON.
- Designing an LLM-based semantic comparison system.
Outcome
Prototypes that automate collection, structure content into JSON and support semantic comparison within a modular architecture.
Technologies
- Python
- Playwright
- LLaMA
- JSON
- GitLab
A similar challenge?
Facing a similar challenge? Let’s talk.
Let’s discuss your data, processes and expected outcome to define an engagement that fits your context.