Back to home
Writing
Notes on shipping production AI
Writing about RAG, fine-tuning, LLM architecture, and the operational work that keeps AI systems running in production.
Contract Advisor RAG: building a high-precision legal expert LLM app
Feb 2024
How I pushed a contract Q&A RAG system to 87% relevance with semantic chunking, hybrid retrieval, and a real evaluation loop instead of vibes.
RAGLLMsEvaluation
Fine-tuning Llama 2 for Amharic embeddings and text generation
May 2024
Enabling quality embeddings and generation for a low-resource language, then wiring the model into a RAG-based Amharic ad-copy builder.
Fine-tuningLLMsAmharic
Automated storyboard synthesis for digital advertising
Feb 2024
Combining EDA on creative assets, YOLO object detection, and image generation to cut storyboard production from a week to a few hours.
Computer VisionGenerative AIAd-Tech
Revolutionizing data analysis with a Redash chatbot add-on
Jan 2024
An agentic RAG add-on that turns natural-language questions into SQL against existing Redash dashboards, so users chat with their data.
Agentic RAGLLMsAnalytics