AI systems for founder-led companiesVol. MMXXVIBangkok

Make your team AI-nativeand finally see the returns.

Everyone has the same models. The returns come from knowing where AI fits in your business and building it well. I help founder-led companies work that out and build it.

Taking on 3 new builds this quarter
Arun Thiru · BangkokI was a staff data scientist at eyeo, the company behind Adblock and Adblock Plus. I worked on products used by hundreds of millions of people.

Trusted by founders, operators and engineers from

Stripe Coinbase JPMorgan IBM Autodesk Aurora Starling Bank Niagen Biotech Stream UK Matcha Growth
01The problem

Everyone seems to be getting results from AI. Your company feels stuck.

It's not the tools. You have the same models as everyone else. The returns come from knowing where AI fits in your business and building it well. That's how a team stops falling behind and starts becoming AI-native. Here's what that can look like inside your company:

Research Systems
Agents that gather research from your own data and the web, and give your team answers they can act on.
AI Employees
Agents that work inside your tools and handle part of the job, with a person checking the output.
Automated Internal Reporting
The reports your team rebuilds by hand each week, built once so they run on a schedule.
Second Brains and Knowledge Bases
A shared store of how your business works, so every other system can build on it.
Fine Tuning Models
Trained on your field and the way you write, so it fits how you work.
02Ways to work together

Do you need me to look, or to build?

Every engagement starts small. Book a paid working session to shape a solution, or a free call to see whether there's a larger project worth scoping.

01

Consulting call.

$500 · 60 min

We review your problem, your current tools, and the bottleneck you want to fix, then sketch a solution together. You leave with a written report that covers the problem as I understand it, a design for the solution, and the tools I would build it with.

Book the $500 call
02

AI opportunity sprint.

Scoped after the call

This is for companies that know AI should help somewhere but need a senior technical partner to find the first workflow worth building. I map how the work happens now, look at the data, the users, and the risks, decide what success means, and pick what to build first.

Book a discovery callfree · 20 min
03

AI implementation.

By project

This is for companies ready to move from talking about AI to a working system inside the business. I build the system, connect it to your tools, check the outputs, put it into daily use, and train your team with documentation they keep.

Talk through a buildfree · 20 min
04

Fractional Head of AI.

Monthly

This is for companies that need ongoing AI leadership before they hire a full team. I own the roadmap, set priorities, decide what to build and what to buy, set vendor standards, help your team adopt it, and report to your executives.

Book a growth callfree · 20 min
03The difference

David vs Goliath.
Why small teams win.

My bet is on small, efficient teams powered by AI.

  • You can quickly test solutions
  • Less bloat, less meetings
  • Less layers between you and the work
 
Working with me
The traditional path
Team
Small core build team
Large team, many handoffs
First working system
In weeks
After a quarter of discovery
Who builds it
The person who advised you
Someone separated from the strategy
When you test
One workflow, live and early
After the deck is signed off
Security, legal, data
Reviewed at defined checkpoints
Bolted on late
What you're left with
A working system your team understands
A strategy deck and a dependency
04Proof

I build with AI every day, including on my own products.

Hundreds of millions of users on Adblock and Adblock Plus, which I worked on as a staff data scientist at eyeo.
$2K in 4 weeks The first version of Helios made $2,000 in its first four weeks, before I built it into a full platform.
Same lean approach I try everything on my own products first. I keep teams small, build one workflow at a time, and make systems you can read and check.

Arun made AI feel simple and gave us the confidence to implement it.

A client. Named case studies are coming.

You should finish more capable, not more dependent.

A good AI consultant makes your team stronger. You should leave with a working system and a clear understanding of how it works, where it can fail, and how to improve it. You should not be left with something only I can maintain.

05Before you book

Questions I get asked.

What does the $500 call actually get me?+
One working session where we review the workflow, the tools, and the bottleneck, then sketch a solution together. Afterwards you get a written report with the problem as I understand it, a design for the solution, and the tools I would use. It's useful whether or not we work together after.
What does "weeks, not quarters" really mean?+
It means we build one workflow at a time, so you can test a working system in weeks and grow from what you learn. It does not mean every project finishes in days. Complex or heavily regulated work still takes the time it needs. The speed comes from a narrow scope, not from skipping careful work.
What if we're not even sure AI is the answer?+
That's a good reason to start with the call or a free discovery call. Part of my job is telling you honestly where AI will and will not help. If a workflow is mostly noise, no tool will fix it, and I'll tell you.
Who actually does the work?+
I do the main build. For larger projects I bring in a small team, and the person advising you stays close to the person building. Security, legal, and domain experts review the work at set points.
What about our data and security?+
People decide on data access, security, and how systems connect. We review those choices at set points and do not skip them to move fast. Before anything goes live, we agree on what data the system needs, how it's handled, and who signs off.
Do you work fixed-price or ongoing?+
Both. Many clients start with the consulting call or a scoped sprint, then move into a build project. Some continue with fractional AI leadership once there's a roadmap worth owning. You grow from results, not from a long commitment up front.

Let's talk about
your workflow.

If your team has a workflow everyone knows is painful, but nobody has turned it into a working system, that's where I fit. I read every message myself and reply the same business day.

3 build spots open this quarter