AI
CompletedEnterprise AI Platform
A private AI platform with local models and external APIs side by side, giving teams a controlled place to use AI with their own documents.
Answer returned5/5
The result comes back, and the data stayed in your control.
The problem
People want to use AI at work, but pasting company documents into public chatbots is a data protection problem.
A business needs one controlled entry point, with known models, known users and data that stays under its own control.
My contribution
- Designed and deployed a self-hosted AI workspace.
- Connected local models alongside external APIs.
- Set up document upload and retrieval (RAG) for questions over internal knowledge.
- Put user accounts, monitoring and backups in place.
Technical architecture
- Interface
- Self-hosted AI workspace with user accounts
- Models
- Local models alongside external APIs
- Knowledge
- Document retrieval (RAG)
- Delivery
- Containers with automatic recovery and backups
Technologies
- Open WebUI
- Ollama
- RAG
- Docker
Highlights
- Local and hosted models in one workspace
- Company data stays under company control
- Backups and monitoring included
Details are generalized to protect confidential information. Company systems, customers, figures, source code and screens are not shown.