When the SDK Says No: Vision-Aware RAG via a 300-Line Python Bridge

📚 Series Building a Knowledge Management App with Oracle AI Database 26ai Part 3 of 3 View all posts in this series → This is Part 3 — the final chapter in turning a text-only knowledge base into a fully vision-aware RAG system After moving CLIP embeddings into Oracle AI Database 26ai, I had something great: true multimodal search. Text queries found images. Images found related text. No OCR, no captions—just CLIP doing its magic in the database. ...

November 22, 2025 · 14 min · Brian Hengen

Three LLMs, One App: Balancing Speed, Privacy, and Power

📚 Series Building a Knowledge Management App with Oracle AI Database 26ai Part 1 of 3 View all posts in this series → I spent a weekend fine-tuning a model for my knowledge management app, designed to handle notes, PDFs, and presentations with Oracle Database 23ai’s vector search (see my management AI post). It aced testing on my RTX 5090 server, but on my M2 MacBook Pro? Barely usable. A query like “Summarize last week’s customer meetings and identify risks” took over a minute, leaving me staring at a spinning wheel while my coffee got cold. ...

October 28, 2025 · 6 min · Brian Hengen

CLIP Inside Oracle AI Database 26ai: Fast, Multimodal RAG

📚 Series Building a Knowledge Management App with Oracle AI Database 26ai Part 2 of 3 View all posts in this series → After the 3-way LLM toggle went live, I turned my attention to embeddings - the invisible glue that powers search and RAG. Oracle OCI GenAI’s Cohere endpoint had been rock-solid in my testing: fast, reliable, and gave me 80 K token context. But every chunk still meant a network round-trip, and images were stuck behind OCR, so text-only embeddings meant photos, diagrams, and whiteboards were blind spots in my knowledge base. ...

November 11, 2024 · 11 min · Brian Hengen

Subscribe to New Posts

Get notified when I publish new AI experiments and research findings.