Embeddings in RAG: The Core of Semantic SearchRetrieval-Augmented Generation (RAG) systems rely on the core idea of converting text into vectors so we can search by meaning instead of keywords. These vectors are called embeddings, and they are thApr 22 · 6 min readblog.chapainaashish.com.np
Reasoning Patterns in LLM: CoT & ReActA direct prompt works well with LLM for a simple task, but when a task needs multiple steps to arrive at a particular solution, a direct prompt is likely to fail because of LLM hallucination and assum3Apr 13 · 9 min readblog.chapainaashish.com.np
Validation & Testing LLM OutputsThe moment we interact with LLMs, we get probabilistic output. They'll return "price": "$45.99" one time and "price": 45.99 the next. Sometimes, they even forget required fields. This might not look lApr 3 · 4 min readblog.chapainaashish.com.np
Observability in AI applicationObservability is an essential topic in modern software development. It is particularly useful when you move the application to the production. Observability in AI application is more extended than traditional application because you need to observe t...Nov 26, 2025 · 5 min readblog.chapainaashish.com.np
Think Before You Call LLM APIOverview The first way to interact with any LLM is through the API in development. This is pretty basic which you will master easily but you can miss some critical aspect which I will cover in this post. Part 1: Understand Tokens Before Getting Excit...Nov 7, 2025 · 7 min readblog.chapainaashish.com.np
