AArunSaravanen
HomeBlogAskAboutNowSetupResume
- AS
Models & agents

AI.

4 entries filed under AI, newest first.

All writing
  • May 8, 20267 min
    Neural Networks 101: How a Computer Learns to Read Your Handwriting

    A simple, visual walkthrough of how neural networks recognize handwritten digits, inspired by 3Blue1Brown's classic Deep Learning series.

    neural-networks
  • April 30, 20263 min
    RAG That Doesn't Lie: Grounding Patterns I Trust

    Most RAG systems hallucinate when the retriever fails silently. Here are the four grounding patterns I lean on in production and the one anti-pattern I keep rem…

    RAG
  • April 15, 20263 min
    Building Agent Loops That Don't Sprawl

    A short, opinionated checklist for keeping an agent loop bounded — tool budget, termination conditions, and the single observability hook that's saved me the mo…

    Agents
  • March 30, 20263 min
    Evals as Product Feedback: From Vibes to Numbers

    How I shifted an AI feature team from 'it feels better' to a small, durable eval set that actually informs product decisions — without turning into another dash…

    Evals
Arun Signature

© 2026 Arun Saravanen · Notes from the work · RSS