Make RAG Trustworthy in Production

Two weeks after launch, a policy assistant confidently told employees they could expense $65/day . The absolute limit was $75 , updated three months earlier. That wasn't a “model hallucination"—it was stale retrieval with no uncertainty handling , quietly shipping bad guidance at scale. This post shows how to stop that. We'll turn "embed → top-k → paste” into a production-ready loop that adapts its strategy, checks its own evidence, and admits uncertainty when support is wea

Nandita Krishnan

Oct 135 min read

FEATURED POST

Never Miss a New Post.

RECENT BLOGS

Make RAG Trustworthy in Production

Two weeks after launch, a policy assistant confidently told employees they could expense $65/day . The absolute limit was $75 , updated three months earlier. That wasn't a “model hallucination"—it was stale retrieval with no uncertainty handling , quietly shipping bad guidance at scale. This post shows how to stop that. We'll turn "embed → top-k → paste” into a production-ready loop that adapts its strategy, checks its own evidence, and admits uncertainty when support is wea

Oct 135 min read

A visual, text-driven timeline explaining the evolution of Retrieval-Augmented Generation (RAG) systems — from the original 2020 “retrieve-then-generate” method to modern agentic and adaptive frameworks. Highlights milestones like HyDE, RAG-Fusion, Self-RAG, CRAG, GraphRAG, and Adaptive RAG, with simple flowcharts and concise explanations showing how retrieval, reflection, and decision-making evolved in AI systems.

The RAG Reformation

From “retrieve-then-generate” to self-reflective, decision-driven systems — this short, illustrated history of Retrieval-Augmented Generation (RAG) traces how retrieval evolved from a simple lookup step into a dynamic, agentic layer of reasoning. Follow the milestones from Basic RAG to Self-RAG, GraphRAG, and Adaptive RAG, and see how modern AI learns when to retrieve, what to keep, and how to revise.

Sep 13 min read

Welcome to My Blog

Hey! I'm Nandita. I've been working in AI and security for several years, building systems and solving complex problems. I'm currently...

May 261 min read

AI Interviewers: Fad or the Future? Why Practicing with Simulated Scenarios Works

Artificial Intelligence is shaking up how companies hire these days. As an engineer, should you worry? We saw a rapid extinction of...

Apr 154 min read

ABOUT ME

Nandita Krishnan is an engineering leader turned founder with over a decade of experience building scalable systems and leading high-performing teams across tech, media, and identity platforms. She’s passionate about career growth, leadership, and building human-centered products.