Agents, evals and developer tools — built in public, explained on Coding Adda.
More about me0K
developers learning on Coding Adda · live
0%
less dev time from the agents & skills I built at Synopsys
0
products shipped in public — CLI, macOS, web
0
published book on prompt engineering
01 — Selected work
Things I've shipped.
AdiClaw
An autonomous CLI coding agent that never touches your disk blind. Every change is staged in a sandbox, verified against your tests, then offered as a patch.
02 — Right now
What I'm up to.
03 — How I build
Determinism over stochastics.
Agents that edit files directly cause silent regressions. Every mutation is isolated in a staging environment, verified against tests, then committed.
Rebuild it from scratch.
I understand databases, KV-cache indexing and multi-agent coordination by re-implementing their core algorithms in plain code.
If you can't eval it, don't ship it.
Prompts and system instructions are software artifacts. Reproducible benchmarks, synthetic eval harnesses and regression gates, never anecdotal runs.
04 — Learn with me
Pick your path.
Notes, videos and teardowns, organised by where you are — not by when I published them.
New to GenAI
Introduction To Generative AI · Basics of Prompt Engineering · Coding Adda GenAI videos
1 note · 1 book · 564+ videos
Write better software
DRY · KISS · YAGNI · Singleton · Builder · State · Strategy
7 notes · interview-ready
Build real systems
Build X From Scratch · paper breakdowns · agent architecture essays
guides · papershelf · essays
05 — The book
Basics of Prompt Engineering.
A practical guide from intuitive prompting to production-grade systems: context windows, structured output, tree-of-thoughts and evaluation gates.
Get it on Amazon →
06 — Let's talk
Collaborations on AI agents and developer tools, talks and workshops, mentoring, or an interesting role — pick whichever door fits.
Open to collaborations · Bengaluru, India