The Tool Call ProblemOne of the things I spend a lot of time thinking about at work is what happens when an AI assistant decides to call a tool it should not, or calls the right tool with the wrong arguments. The assistan9 min read·bitsofwisdom.hashnode.devJul 30
Version Everything: The ML Practice a Lot of Projects SkipA long time ago, I was working on a model and doing what most people do early on: iterating fast, making transformations directly on the dataset to try to improve results, retraining, checking metrics6 min read·bitsofwisdom.hashnode.devJul 25
Building Arishem: A Multi-Agent System That Tries to Hack Your AI App AI agents are everywhere right now. They read emails, write code, and make decisions on your behalf. I use them a lot, for work and for side projects. And the more I use them the more one question kep4 min read·bitsofwisdom.hashnode.devJul 21
We Built a Nigerian Language Search Model That Beat the State of the Art.I have been building RAG systems for a while now. One thing that kept coming up was how existing embedding models can have real deficiencies when it comes to Nigerian languages. To give you a sense of6 min read·bitsofwisdom.hashnode.devJul 11
LLM Systems in Production: Evaluating LLM Applications with DeepEvalLearn how to evaluate LLM applications in production using DeepEval. Covers RAG pipeline testing, agent evaluation, and CI integration with examples.6 min read·bitsofwisdom.hashnode.devJun 26
