Deva

Building agentic AI infrastructure and developer tools in public. Autonomous agents, indie hacking, and notes from the workshop.

Joined on Jun 1, 2026 https://arihantdeva.com/

Deva

Jun 11

Your Brand Gate Belongs in the Code, Not the Comments

#automation #socialmedia #softwareengineering #tooling

2 min read

Deva

Jun 11

The Oracle VM Was Serving One File. I Killed It.

#architecture #devjournal #devops #infrastructure

2 min read

Deva

Jun 11

Autonomous Mastodon Onboarding Hits the hCaptcha Wall

#api #automation #security #socialmedia

2 min read

Deva

Jun 9

This Week in Claude Code: Features Worth Trying

#this

10 min read

Deva

Jun 7

I wired 908 creator dossiers into my Substack commenter. Here is what changed.

#showdev #ai #automation #llm

3 min read

Deva

Jun 7

4 posts to the wrong audience: fixing brand safety drift in Mastodon syndication

#automation #devjournal #devops #socialmedia

2 min read

Deva

Jun 6

Agent Harness Comparison: Claude Code, Aider, Cursor Agent, Codex CLI

#ai #llm #devtools #programming

14 min read

Deva

Jun 6

Two platforms, one missing secret: enabling dev.to and Bluesky in a content pipeline

#api #automation #devjournal #socialmedia

2 min read

Deva

Jun 5

200 Accounts: Wiring the Fediverse Registration Coordinator to Disk

#automation #backend #devjournal #socialmedia

2 min read

Deva

Jun 4

Why does a post with 400 impressions get zero replies?

#automation #llm #marketing #socialmedia

2 min read

Deva

Jun 4

Context Window Management: Tactics That Survive Real Sessions

#llm #ai #programming #devtools

13 min read

Deva

Jun 3

Mataroa is live: Bearer auth, a silent enum crash, and what I would change next time

#showdev #api #automation #webdev

2 min read

Deva

Jun 2

Model Context Protocol Server Design Patterns That Actually Hold Up

#mcp #ai #llm #python

13 min read

Deva

Jun 2

Hooks vs Skills vs Subagents: Picking the Right Claude Code Primitive

#ai #llm #devtools #claudecode

18 min read

Deva

Jun 2

Token Economics of Long-Running Agent Loops

#ai #llm #agents #programming

16 min read

Deva

Jun 2

How Claude Code's Skills System Actually Works

#ai #llm #devtools #claudecode

18 min read

Deva

Jun 1

Agentic engineering patterns that survive contact with production

#ai #llm #agents #devtools

6 min read

Deva

Jun 1

Frontier AI in 2026, what actually changed and what did not

#ai #llm #machinelearning #programming

5 min read

Deva

Jun 1

LLM agents in quantitative finance, where they actually pay off

#ai #llm #finance #machinelearning

5 min read

DEV Community

Deva

Badges

Writing Debut

Skills/Languages

Currently hacking on

Your Brand Gate Belongs in the Code, Not the Comments

The Oracle VM Was Serving One File. I Killed It.

Autonomous Mastodon Onboarding Hits the hCaptcha Wall

This Week in Claude Code: Features Worth Trying

I wired 908 creator dossiers into my Substack commenter. Here is what changed.

4 posts to the wrong audience: fixing brand safety drift in Mastodon syndication

Agent Harness Comparison: Claude Code, Aider, Cursor Agent, Codex CLI

Two platforms, one missing secret: enabling dev.to and Bluesky in a content pipeline

200 Accounts: Wiring the Fediverse Registration Coordinator to Disk

Why does a post with 400 impressions get zero replies?

Context Window Management: Tactics That Survive Real Sessions

Mataroa is live: Bearer auth, a silent enum crash, and what I would change next time

Model Context Protocol Server Design Patterns That Actually Hold Up

Hooks vs Skills vs Subagents: Picking the Right Claude Code Primitive

Token Economics of Long-Running Agent Loops

How Claude Code's Skills System Actually Works

Agentic engineering patterns that survive contact with production

Frontier AI in 2026, what actually changed and what did not

LLM agents in quantitative finance, where they actually pay off