Ian Macabulos
Full-Stack Developer
Full-Stack Developer
KodaSync is a professional intelligence hub designed to centralize technical knowledge for software engineers. It solves the "lost context" problem in modern development by utilizing high-dimensional vector embeddings to understand the underlying intent of your code. It combines a high-performance Monaco Editor with contextual AI agents to transform scattered snippets into a searchable, neural knowledge base.
Experience the neural research assistant live or inspect the architecture on GitHub.
An AI workspace using Retrieval Augmented Generation (RAG) to provide project-specific answers. It automatically retrieves relevant notes from your library to provide context-aware solutions.
A full-featured code editor built with the Monaco Editor engine (VS Code core). It includes AI actions to Scan Security, Generate Tests, and Document code instantly.
Uses FastEmbed (384-dimension vectors) to enable natural language search. Includes a Knowledge Scraper to import technical documentation directly from URLs.
The main challenge was orchestrating the "Hybrid Architecture" between a modern Next.js frontend and a high-performance Python backend. I needed to ensure that heavy vector processing and AI inference didn't block the UI, while maintaining a real-time editing experience in the Monaco Editor.
To ensure the neural engine remains reliable, I implemented a comprehensive testing strategy that validates both the API logic and the AI's streaming capabilities:
Using Pytest and TestClient, I wrote tests covering auth flows, vector search, and RAG logic. I specifically mocked AI streams to validate chat responses without calling expensive external APIs.
A robust GitHub Actions workflow that spins up ephemeral Docker containers for PostgreSQL (pgvector) and Redis. This ensures the full integration suite runs successfully on every push before deployment.
KodaSync pushed my boundaries in Full-Stack Engineering, specifically in bridging the gap between application logic and AI infrastructure. Mastering Docker for containerization and implementing a real-time RAG pipeline gave me the tools to build production-grade AI applications.