Skip to content
Andrés

Andrés Ramírez Sierra

Full-time: Founder & CEO, RootBlocks.

Intelligence is now artificial,
accountability remains human.

I build with intelligence, data, and processes, and I build them fast & auditable; the kind of software even regulated industries can actually trust.

Let's work together Explore SliceOps

The method

Achieved with my framework

SliceOps, the open framework I authored. One builder, since May 2026, every decision on the record.

124x
faster than traditional building →
1,000+
decisions on the record
15
agents in parallel
6
systems built & measured

The outcomes

What I've put into production

In production across regulated markets, serving real people at real scale.

1M+
people served
(unique individuals · as of Jul 2026)
5,400+
companies served
54M+
data points managed

What I believe

Everything I build comes from a few convictions. They're the edge, the part no tool can copy.

  • 01People follow people, not algorithms.
  • 02People should own their data.
  • 03Intelligence is a co-architect, not a replacement.
  • 04Non-technical doesn't mean non-builder.
  • 05Simplicity is the highest discipline.

The story

Years of building taught me less than one failure did.

In 2019 I lost a company I'd spent years building: a lending technology that Experian distributed across Colombia, and that put more than $5 billion of credit into the hands of millions of people. There were forces outside my control. But I was the one steering, and I made calls that left it fragile. It ran out of cash, and I lost it. With my reputation on the line, the most useful thing I did was stop blaming the storm and own the ship.

What I got wrong: I'd built a bet, not a system. Something that only survived while conditions held. So I rebuilt around what I should have done from the start.

Everything I build now comes from that. The failure wasn't bad luck I survived; it was the education that taught me to build things that don't break. And to own it when they do.

Why all of this

As machines do more of the work, the real risk isn't that AI gets too smart. It's that people end up owning less, controlling less, understanding less. Opaque models. Data you don't hold. Decisions no one can trace. I build the opposite. Every company I start is the same idea in a different layer: intelligence that stays accountable, data that stays yours, processes that keep a human in charge. And I build it where it's hardest, in regulated industries, because what survives there works anywhere.

How I build

Every company runs on SliceOps, the open framework I wrote: decision-first, with AI as a co-architect and every choice on the record. It is the difference between building a bet and building a system, and it is why a new product now takes days instead of years.

The method behind the numbers

I wrote the framework I needed.

SliceOps is the open, decision-first framework and audit plane for AI-first software engineering, so multi-agent teams ship auditable software, not vibe code. It's how a small core ships across the whole ecosystem, with every decision on the record.

See SliceOps The speed it buys →

What one on the record looks like

Envelope-encrypt the personal data vault (DEK / KEK)

Context
Personal data must be encrypted at rest with per-record keys, recoverable, and auditable.
Decision
Adopt envelope encryption: a per-record data key wrapped by a rotating key-encryption key held in the KMS.
Consequence
Key rotation never re-encrypts payloads, and every access is logged and traceable.

What I build

One path, in order: the company I lost, the one I run, the framework that changed how I build, the company I'm building on it, and what I'm building to prove it works. Three layers run through all of it: intelligence, data, and processes, each accountable to a human. I built the first the old way, over years. What runs on the framework now takes days. See the before and after.

Before

Banlinea

Processes

2011 to 2019, my first company. A lending technology Experian distributed across Colombia. I lost it in 2019, and it became the education behind everything since.

Active

Insttantt

Processes

The company I built next, alive and in production, fully bootstrapped. Regulated services run online end to end, now run by its own team.

Visit Insttantt

The AI Paradigm Shift

SliceOps

Intelligence

The real problems never went away: software is expensive, teams are hard to orchestrate, hiring is hard, and good code is harder still. AI changes everything, and hits the same problems, only now at exponential scale and speed. So I wrote the methodology I needed: SliceOps, the open framework where every decision is recorded by construction. The foundation the rest stands on.

Explore SliceOps

Building now

RootBlocks

Intelligence

The company I'm building on SliceOps: governed agentic software construction, so any enterprise can build with agents without losing control or compliance.

Every AI-written change ships with proof.

Visit RootBlocks

Testing what it can do

Datta

Data

Built to push the approach to its limits: a data account that gives AI agents verified context about a person or a company, their goals, decisions, and credentials, so they act with real judgment instead of generic advice. Pre-launch.

Resolvi

Processes

The same, in customer service: full structured context of your company and your customer, never a black box.

Let's work together.

Three ways we can build something real.

Use the tools

SliceOps is open, adopt it free. RootBlocks brings that governance to your enterprise. Datta is on the way. Take what I've built and run with it.

Partner

I take on a small number of ventures and open-source collaborations at the frontier of intelligence, data, and processes. If you're building something hard, let's talk.

AI for your company

Built, not advised. Governed, auditable AI in production, made to last. No decks.

Want this for your team? That's RootBlocks

Send a message

Tell me what you're building. I read everything that comes in.

Book a call

Prefer to talk? Grab a time that fits, a short intro or a deeper look at the tech.

See available times →
GitHub, @anrasiLinkedIn