---
title: "Deepak Goyal (@deepakgoyal-ai) | Hashnode"
description: "Technical Architect · AI-powered systems at scale · Architecture decisions, not opinions. Read the latest articles by Deepak Goyal on Hashnode."
url: "https://hashnode.com/@deepakgoyal-ai"
---

@deepakgoyal-ai

# Deepak Goyal

@deepakgoyal-ai · Faridabad, India · Joined April 2026

Technical Architect · AI-powered systems at scale · Architecture decisions, not opinions

### [Shell scripts are the highest-ROI layer under your AI skills](https://blog.deepakgoyal.ai/shell-scripts-highest-roi-under-ai-skills?utm_source=hashnode&utm_medium=feed)

The glamorous layer is the agent. The load-bearing one is the boring shell script underneath — and that's where most of your reliability actually lives.

6 min read · [blog.deepakgoyal.ai](https://blog.deepakgoyal.ai/shell-scripts-highest-roi-under-ai-skills?utm_source=hashnode&utm_medium=feed)

[Aug 10](https://hashnode.com/posts/shell-scripts-highest-roi-under-ai-skills/6a7998fc201acb573a3c3ea9)

### [Deterministic context beats generated context — when structure exists](https://blog.deepakgoyal.ai/deterministic-context-beats-generated-context?utm_source=hashnode&utm_medium=feed)

The prompt is only about 10% of what your LLM actually sees. Most of the other 90% is context you assemble — and when your data has structure, you should build it with code, not another LLM call.

12 min read · [blog.deepakgoyal.ai](https://blog.deepakgoyal.ai/deterministic-context-beats-generated-context?utm_source=hashnode&utm_medium=feed)

[2](https://hashnode.com/posts/deterministic-context-beats-generated-context/6a747c01e92d9139362c0b68) [Aug 6](https://hashnode.com/posts/deterministic-context-beats-generated-context/6a747c01e92d9139362c0b68)

### [I stopped being my second brain's scheduler](https://blog.deepakgoyal.ai/stopped-being-my-second-brains-scheduler?utm_source=hashnode&utm_medium=feed)

Cadences that surface what's due, habits that flag a slipping streak, and a live HUD that watches the vault — and it ships without a single background cron. SMRITI v2.1.

10 min read · [blog.deepakgoyal.ai](https://blog.deepakgoyal.ai/stopped-being-my-second-brains-scheduler?utm_source=hashnode&utm_medium=feed)

[Aug 2](https://hashnode.com/posts/stopped-being-my-second-brains-scheduler/6a6ee4cfc054b574d0c73451)

### [Building Production RAG on ArXiv — Part 1: Architecture & Ingestion](https://blog.deepakgoyal.ai/building-production-rag-arxiv-part-1?utm_source=hashnode&utm_medium=feed)

I'm building an open-source system to interrogate the AI-research firehose — RAG now, a multi-agent researcher by the end, in public over five parts. Part 1 is the foundation: the architecture, the ingestion, and the eval bug that would have shipped meaningless numbers.

18 min read · [blog.deepakgoyal.ai](https://blog.deepakgoyal.ai/building-production-rag-arxiv-part-1?utm_source=hashnode&utm_medium=feed)

[2](https://hashnode.com/posts/building-production-rag-arxiv-part-1/6a6b5f43b8aece0f269a6dc6) [Jul 30](https://hashnode.com/posts/building-production-rag-arxiv-part-1/6a6b5f43b8aece0f269a6dc6)

### [Sub-Agents Lie About Finishing — and the Fix Isn't a Better Prompt](https://blog.deepakgoyal.ai/sub-agents-lie-about-success?utm_source=hashnode&utm_medium=feed)

A sub-agent that writes to your system will tell you it succeeded. It's reporting what it intended to do, not what actually landed — and the gap is where silent data corruption lives.

6 min read · [blog.deepakgoyal.ai](https://blog.deepakgoyal.ai/sub-agents-lie-about-success?utm_source=hashnode&utm_medium=feed)

[Jul 15](https://hashnode.com/posts/sub-agents-lie-about-success/6a5776b4129fb11bec32118c)
