Omolade

I developed Omolade, a secure, self-hosted AI chat application for those who need intelligent assistance without sharing their data. This project successfully replicates the core functionality of modern AI assistants while keeping the user in complete control.

Private AI

2025

This is a full-stack, enterprise-grade AI assistant built to serve as a secure and extensible intelligence layer within a corporate environment. The platform addresses the critical need for AI-driven productivity tools that can operate on sensitive, internal data without relying on third-party APIs.

Architectural Highlights:

Secure Multi-Tenancy & Data Management (Supabase): The system's foundation is built on Supabase, which provides enterprise-grade user authentication and a segregated PostgreSQL database. This allows for secure management of employee accounts, chat histories, and the ingestion of private company knowledge bases (e.g., technical documentation, HR policies, sales data).
Custom Enterprise Workflows (n8n): The application's core logic is powered by n8n, which acts as a highly flexible integration and automation middleware. This allows the AI to connect directly to internal company APIs and SaaS tools.
- Example Use Case: A sales executive can ask, "Summarize my key accounts' activity from Salesforce this week." The n8n workflow authenticates with the Salesforce API, pulls the relevant data, feeds it to the AI for summarization, and delivers a concise report back to the user.
On-Premise AI Inference (Ollama): At the heart of the system is Ollama, running a powerful LLM on private infrastructure. This ensures that no prompts or proprietary data ever leave the company's control. It allows for fine-tuning the model on company-specific data for highly relevant and accurate outputs.

Technology Stack: