I'm Haroon Choudery. I'm the founder of Seeko.
I've spent the last several years building AI infrastructure. I built Autoblocks, one of the first evaluation and observability platforms for AI systems. Engineering teams at companies like Hinge Health and ClickHouse use it to make their agent systems reliable in production.
I write AI Ready, a newsletter on AI readiness that goes out to 50,000+ subscribers. And I host AI Ready, a show where I talk with engineering leaders at companies like OpenAI, Intercom, and Retool about how they actually build with AI.
Before all of that: ML consulting at Deloitte, early team at Hex, co-founder of AI For Anyone.
Why Seeko
I started Seeko because I kept having the same conversation.
Engineering leaders would reach out after reading the newsletter or hearing me on a podcast. They'd say something like: "We're using Copilot, we've built a couple internal tools, but we don't really have a strategy. We're not sure what to build next or how to think about this."
They didn't need a training workshop. They didn't need a 200-slide deck from a consulting firm that's never shipped a model. They needed someone who'd actually been in the room, building this stuff, to sit with their team and figure it out together.
That's what Seeko does.
We help engineering, operations, and product teams figure out where AI creates real leverage, then build it. Strategy when you need direction. Hands-on implementation when you need to ship. Advisory when you need someone to call.
Seeko is a small, hands-on practice. Depending on the engagement, I work directly with your team or bring in specialists from my network. Either way, I'm involved in every project. You get senior attention from day one, not a junior associate learning on your codebase.
The thesis
There's an idea that runs through everything we do at Seeko.
The best AI systems don't stay experimental forever. They start flexible: exploring, figuring out the right workflow. Over time, the stable patterns crystallize into reliable, fast, deterministic code. The AI handles what's novel. Everything else hardens into infrastructure your team owns.
Most companies are somewhere on this spectrum, and almost all of them think they're further along than they actually are. That gap between perception and reality is where the most important work happens.
We write about this regularly in the newsletter and on the blog.
AI Ready
AI Ready is where most people find me.
It's a newsletter about AI readiness: how teams are actually building with AI, what's working, what isn't, and what the patterns look like when you zoom out. It's not news roundups or product announcements. It's the stuff I'm learning from working with teams and building systems myself.
50,000+ people subscribe. If you want to understand how I think about this space, that's the best place to start.
Get in touch
If you're working on something where AI should be playing a bigger role, and you're not sure where to start or what to do next, tell us what you're working on (or just reply to any newsletter issue). I read everything.