The Build · Harness Commander

From the boardroom to the build.

An autonomous AI operating system. Built in 30 days — from first principles.

Nine years running enterprise software from the customer's side — $2M+ ARR, India's biggest brands, not a line of code among them. Then he stopped advising on AI and started building it. Thirty days later, he'd shipped an autonomous harness that plugs into any workflow and gets smarter every session — arriving, from first principles, at the same patterns the research world was just naming: context engineering, layered memory, compound systems.

AI OS in 30 days built on 5 research ideas $2M+ ARR background Top 1–2% globally
scroll
The Arc

Nine years in the making.

Nobody builds an autonomous AI system in a month by accident. The leap only looks sudden — here's the road that made it inevitable.

The foundation
A researcher first
An M.Tech in Manufacturing Science, a real love of material science, a published paper — and a serious thought about a PhD. He didn't take it. But the researcher's instinct stayed: go deep, reason from first principles, trust the problem over the playbook.
First-principles roots
2019
He saw it early
Handed the AI Summit to produce, he did what a researcher does — went deep. Mapped the global AI landscape before it was mainstream, spoke with leaders across the field, and watched Dario Amodei leave OpenAI because safety had to come first. The conviction started here — years before the hype.
AI conviction · since 2019
The decade
At the top table
Conversational AI at Verloop and Haptik. Martech at MoEngage. Across the decade he earned the trust of CXOs and ran India's biggest brands — Swiggy, Ola, Reliance, Amazon miniTV — across a $2M+ ARR portfolio. He knew exactly what enterprises needed from AI — he just hadn't built it himself yet.
Enterprise · CXO · $2M ARR
The turn
The reset
He stepped away — no script, no safety net. And discovered what Claude could really do: not a chatbot, an operating layer. Something that could think alongside you, hold context, build.
The discovery
Now
The call to build
The builder's instinct took over. Harness engineer. Vibe coder. He stopped looking for projects and started solving his own — and out came Harness Commander, Caby, an autonomous video pipeline. The system he'd spent nine years getting ready to build.
Harness engineer · vibe coder
Where It's Going

Not built for one domain.
A harness for any process.

Harness Commander is an intelligence layer that plugs into any workflow — bringing persistent memory, encoded judgment and self-correction — and compounds with every session. Customer Success is simply where it's being proven first.

01
Pluggable Harness
the core
Drop it onto any process — a tool, a team workflow, a whole function. It arrives with memory, skills, hooks and self-healing already wired in. Plug in, and it runs underneath.
02
Compounding Intelligence
the engine
Every session writes intelligence back, not just data. Whatever it's plugged into gets measurably sharper over time — the loop is the product, not any single output.
03
Agentic CS Ops
proven first
The first vertical: renewal intelligence, adoption signals, expansion whitespace, stakeholder mapping — the operating layer that runs underneath every CSM.
04
CS Copilot
the daily driver
QBR narratives, implementation & TAM workflows, and a communication wizard across Slack, Gmail, Jira, Zendesk and Gainsight. AI prepares the play; the human makes the call.

Proven in the domain he knows cold — built to plug into the next one just as easily.

The Build

Not a chatbot. An operating system.

Harness Commander runs underneath everything he does — a self-correcting, compounding intelligence harness, built in thirty days and running in production every day since.

Persistent memory
Survives every session and every context reset. It never forgets who he is or what it's working on.
Encoded skills
Judgment, not saved prompts. One command runs a full multi-step playbook.
A dreaming layer
Synthesises patterns across sessions while he's away — consolidating what matters before the next session.
Self-healing
Detects its own failures and fixes them, without being asked.
Cross-session relay
Nothing built in any session is ever lost.
5-minute recovery
The entire system rebuilds from scratch on any machine, from git, in under five minutes.
Built on the patterns the field was naming in parallel
Claude Dreaming · Anthropic Harness Engineering Layered Memory · CoALA Compound AI · Berkeley Context Engineering · Karpathy & Lütke

The field named them. The edge here was wiring them into one loop that compounds — boot with the right context, work inside structured memory, close with judgment. The next session starts smarter than the last.

Explore the full architecture →

And the OS wasn't the only thing. Caby — an autonomous Telegram bot. A self-running video pipeline. All from the same instinct: stop hunting for projects, solve the real problem in front of you.

The Edge

Anyone can prompt.
These show what he actually is.

Four moments. Click any to see what's behind it.

The Benchmark

Rated by Claude.
Independently.

Claude assessed the harness against a global benchmark of practitioners. No self-reporting. No external panel. The system was audited, scored, benchmarked — and the result surprised even Raj.

Top 1–2% globally

Among Claude practitioners worldwide. This intersection — CS domain depth + AI infrastructure builder — doesn't exist in market.

10.5 / 11 on architecture audit

First-ever run of the system-design-check agent. Caught a regression and fixed it autonomously in the same run.

System Architecture9.5/10
Domain Fusion10/10
First Principles9/10
Originality9.5/10
0/ 10
Performance Rating · June 2026
Top 1–2% · Global
What's Next

Bring harness engineering to enterprise.

The long-term vision is products. But first — join an organisation serious about making AI work at scale, and sharpen harness engineering at enterprise level. Harness Commander is the proof of work. This site is the portfolio.

Building Soon
Agentic CS Ops
An agentic harness for CS Operations.
in build
CSM Mastermind
The CSM's daily co-pilot.
in build
Ecomm Shopping Bot
A conversational shopping assistant.
coming soon
What I'm looking for
→

A role where I implement AI harness engineering at enterprise level — helping teams build reliable, repeatable AI workflows that actually ship

→

Organisations serious about the gap between AI capability and AI reliability in production

→

Teams where CS domain expertise + AI infrastructure thinking = genuine competitive advantage

Get in touch LinkedIn ↗ See the architecture →
chiefcommander.ai · Built with Claude Code · 2026