Kareem Singleton

KAREEM SINGLETON · PHILADELPHIA, PA

Complexity belongs inside the system, not on the person using it.

I build systems that handle more of the repetitive work, preserve what's already been learned, and leave people more room for the part that actually needs them.

Kareem Singleton
AI Systems Architect · Product Builder · Founder, Capsule Foundry

Explore my work Work with me View résumé
The local creation console used to define a development objective.
Kareem Singleton

ABOUT

I built my way in.

Not that long ago, coding looked a little like wizardry to me.

I got into it through AI. First by experimenting with different models and different ways of structuring problems, then by building increasingly complicated systems around them.

Somewhere along the way, I realized I was getting good at something bigger than prompting.

I could take a messy problem, figure out the systems underneath it, break those systems into smaller pieces, and figure out how those pieces needed to work together.

So I started building.

And then I built a system to help me build.

Meet Spiral One

THE SYSTEM BEHIND THE SYSTEMS

Spiral One

The system I built to build systems.

The models keep getting better. That wasn't really my problem anymore.

My problem was everything around them.

Important information gets trapped in a session. Context disappears. A model spends time figuring out something another model already figured out yesterday. You stop halfway through a build and spend part of the next session reconstructing where the hell you were.

And as I got better at building, I kept thinking:

Why isn't the system getting better with me?

So I built Spiral One.

It's the environment I build through now. It helps manage the context, the work, the models, the verification and what we've already learned, so I can spend more of my time on the part I actually care about.

Figuring out what to build next.

The workspace and context surface around a development objective.

SPIRAL ONE ARCHITECTURE

Choose a system layer to explore how it carries the work.

Each layer has a specific job. Together they carry context into working software, verify what it built, learn from it, and recover when things go wrong.

THE LOOP

I want the build to teach the builder.

CouncilBuildVerificationFHRAKnowledgeNext build

Council can expose something we didn't consider. The build tests the idea against reality. Verification checks what actually happened. FHRA steps back and looks at the whole thing. The Empirical Knowledge Ledger keeps what was worth learning.

The next problem shouldn't always start from zero.

Go deeper into Spiral One

BUILT THROUGH SPIRAL ONE

The proof is in what it builds.

Spiral One isn't something I built just to think about building.

I use it.

These projects pushed it into different kinds of problems, from computer vision and native mobile development to local AI, behavioral health tools and portable AI memory.

The apps got better through Spiral One.

Spiral One got better because I built the apps.

01 · BUILT THROUGH SPIRAL ONE

Courier Copilot

A shop-and-deliver offer displayed alongside its calculated advisory.

Some bad offers don't look bad.

SwiftUI · ReplayKit · Vision OCR · Local processing

Courier Copilot overview

You get an offer: $8 for 4.2 miles and 24 minutes.

Is that good?

Now do the math, think about the restaurant, where the trip leaves you, how long you might wait, and whether the whole thing is actually worth it.

Oh, and you might be driving.

Courier Copilot reads the visible offer, does the basic math, and gives me a simple signal while there's still time to decide.

I got tired of doing that.

Open the hood

02 · BUILT THROUGH SPIRAL ONE

Pocket Spiral

Pocket Spiral chat with Qwen3 0.6B / Q4_0 selected and airplane mode visible.

There's a difference between having access to something and owning it.

On-device inference · GGUF · iPhone · Local-first

Pocket Spiral overview

Most AI lives somewhere else.

Pocket Spiral is an experimental AI environment built around language models that can run directly on my phone.

You should be able to pick it up, use it and feel like it's yours.

Open the hood

03 · BUILT THROUGH SPIRAL ONE

Levo

An intensity-based check-in narrows the available next steps.

When thinking harder isn't the answer.

SwiftUI · Jetpack Compose · Local-first · DBT-informed

Levo overview

Sometimes you already know what might help. The problem is remembering it, choosing it and actually doing it when you're already overwhelmed.

Tell it where you are. Get a small set of things you can try. Record what happened. Come back later and see what actually helped.

Sometimes what you need isn't another thing to think about. You need an off-ramp.

Open the hood

04 · BUILT THROUGH SPIRAL ONE

AI Memory Card

Imported conversation records searched through the optional Google Sheets interface.

Your best ideas shouldn't disappear into old chats.

Python · ChatGPT · Claude · Gemini · Local archive

AI Memory Card overview

I use different AI models for different things. That worked really well until I actually wanted the knowledge to move with me.

AI Memory Card turns conversation exports from different AI providers into a searchable, portable archive that stays under the user's control.

Your AI usage should be transferable. Your history should be portable. And you should own it.

Open the hood

DIFFERENT PROBLEMS. SAME INSTINCT.

What can the system handle so the person can spend their intelligence on what actually needs them?

That's the kind of software I want to build.

A NOTE TO THE BUILDER

Let the dreamers keep dreaming.

The weird idea matters. The problem somebody notices because they've actually lived it. The connection their particular brain makes. The half-baked thought at two in the morning that might turn into something real.

Let the system handle more of the repetition, reconstruction and overhead around it.

Then build.

WORK WITH ME

Let's build around the idea.

CapsuleFoundry@gmail.com Connect on LinkedIn
Schedule a discovery call View résumé