Home - Chetan Sroay

AI Engineer — Three Live Products, Built Solo

Chetan Sroay

I build AI products end to end by directing a team of AI coding agents. Three are live — Believele, MarketingSoHigh and Kompense — plus Forge, the internal agent platform that runs their marketing and operations. One engineer, from the first commit to the app stores.

Scroll ↓

01 — About

I don't write most of the code any more — I build three live AI products by directing a team of AI coding agents.

Three live products under Techno Believe Solutions Ltd, plus the agent platform behind them. Kompense (competitive intelligence), MarketingSoHigh (organic marketing) and Believele (an AI job-search assistant on Google Play, the Chrome Web Store and the web). Forge is internal: agents draft my marketing and outreach, I approve anything high-stakes, and anything it cannot verify is held back.

These are live products rather than demos: shipped to Google Play, the Chrome Web Store, WordPress.org and the open web. All three are early-stage, and every number on this site comes from my own systems and is labelled with its source.

Six Microsoft cybersecurity course certificates (Coursera). Autonomous systems have to be secure systems, so Row Level Security, tenant isolation and least-privilege permissions are part of how I build, not an afterthought.

chetan@london ~ zsh

chetan@london:~$ whoami

AI engineer · founder, Techno Believe Solutions Ltd

chetan@london:~$ cat skills.json

{

"ai": ["Claude Code", "MCP servers", "n8n"],

"security": ["Azure", "Defender", "Wireshark"],

"stack": ["React", "Node.js", "Supabase"]

}

chetan@london:~$ echo $STATUS

Open to AI engineering roles — UK, UAE, remote ✓

Available for a permanent role — my products are live and largely automated

02 — Method

How I Work Now

Writing The Code

I typed it

Implementation was the job, and the limit on how much I could ship

I brief it

One agent per codebase, each with a written brief saying what it owns and which code it must never touch

Hover or click →

Data Processing

8 hours

Manual data entry, validation, and reporting across systems

15 minutes

AI agents process, validate, and sync data in real-time

Hover or click →

Publishing

Spot-checked

Agents drafted, I skimmed, and invented claims reached me looking clean

Fail-closed

A fact bank is the only source of first-person claims, and if the checker errors the post is held, not sent

Hover or click →

Measurement

Green dashboards

421 articles, 1.27 million words, every quality gate passing — and 10 clicks, because 40% of our page-one keywords had no search volume

Real-time

Live dashboards with automated insights and alerts

Hover or click →

05 — Tech Stack

Tools & Technologies

#01

n8n

#02

Anthropic Claude

#03

Make.com

#04

Zapier

#05

Claude Code

#06 MCP Servers

#07 API

#08

Python

#09

JavaScript

#0A

React

#0B

Node.js

#0C

Tailwind CSS

#0D

Supabase

#0E

Postgres

#0F

Replit

#10

Azure

#11 Defender

#12

Wireshark

#13

Metasploit

#14

WordPress

#15

GA4

06 — Capabilities

AI Automation

n8n, Make.com, Zapier — complex multi-node workflows at enterprise scale

Intelligent Agents

Claude Code, MCP servers, multi-agent review — design, direct and verify AI agents

Cybersecurity

Azure, Defender, Wireshark, Metasploit — 6 Microsoft certifications

Full-Stack Dev

React, Node.js, Tailwind CSS, Supabase, Postgres — end-to-end applications

API Integration

REST, webhooks, MCP Servers — connecting any system to any system

Development

Python, JavaScript, Replit — rapid prototyping to production code

Microsoft Security, Compliance & Identity Azure Cybersecurity Tools Identity Protection & Governance Cybersecurity Solutions & Defender Cybersecurity Management & Compliance Advanced Cybersecurity & Capstone Microsoft Security, Compliance & Identity Azure Cybersecurity Tools Identity Protection & Governance Cybersecurity Solutions & Defender Cybersecurity Management & Compliance Advanced Cybersecurity & Capstone

Tools & Technologies

The Agentic AI Governance Problem Nobody’s Talking About

GraphRAG: Why Knowledge Graphs Are the Missing Piece in Enterprise AI

Why 90% of AI Automation Projects Fail (And How to Be in the 10%)