YesPress Profile / Founder & CEO
Bruno
Ruyu
The physicist who spent 2005 teaching machines to think - and 20 years building the enterprise tools to prove he was right.
Founder & CEO, Teramot
YesPress Profile / Founder & CEO
The physicist who spent 2005 teaching machines to think - and 20 years building the enterprise tools to prove he was right.
Founder & CEO, Teramot
The Story
Bruno Ruyu learned physics at a glacier-ringed institute in Patagonia. He trained his first neural network the year YouTube launched. He didn't rush.
At Instituto Balseiro - nestled in the Andes, highly selective, Argentina's answer to Caltech - you learn to think before you ship. Ruyu absorbed that discipline and carried it through two decades of energy-sector data work: at YPF (Argentina's state oil company), at Xerox, at Grupo San Cristóbal as Chief Analytics Officer, at Reba as Chief Data Officer. Sixteen-plus years of building pipelines, running simulations, managing uncertainty at scale.
Then in 2022, he did what every long-patient technologist eventually does. He stopped advising companies on their data problems and started solving them at the root.
Teramot is that solution. It connects to your databases, writes production-ready SQL, builds ETL pipelines without code, and - critically - lets you query your entire data stack through ChatGPT, Claude, or any AI assistant via MCP (Model Context Protocol). The pitch is blunt: your data team is expensive, slow, and scaling badly. Teramot is 10x faster, SOC 2 certified, and it doesn't take vacations.
That's not a pitch deck line. It's the result of watching enterprise data engineering break the same way, at the same bottlenecks, for 20 years straight.
"We are building the first Artificial Data Team to empower companies to leverage their internal information - from those with no data strategy to those with massive amounts of data."
- Bruno Ruyu, Founder & CEO, TeramotIn 2005, when most of the world was still figuring out how to use RSS, Bruno Ruyu sat down and trained a neural network. He was a physicist by training - careful about data, allergic to hype, fluent in the language of uncertainty. And he believed, immediately, that what he had just done would change everything.
He was not wrong. He was just early by about 15 years.
That patience - the patience of a scientist who knows the math is right but the infrastructure isn't ready yet - is what separates Ruyu's AI work from most of the venture-backed sprint-and-pivot stories that define the sector. He didn't pivot to AI. He built toward it deliberately, accumulating every credential and scar along the way.
At YPF, Argentina's national oil giant, he ran data and analytics operations for an organization where bad models cost money on a geological timescale. He learned cloud-native architectures, DevOps practices, microservices - not as resume line items but as the toolkit you build when a system actually has to work. He led what he now describes as his first proper "Data Engineering" project in 2014, at a large corporation, where the lessons were expensive and durable.
The Instituto Balseiro in Bariloche is so selective that most Argentines have never heard of it - it accepts roughly 50-80 students per year from a national pool. Alumni include Nobel Prize contributors and leaders across energy, defense, and tech. Ruyu's degree from there isn't just a credential. It's a way of seeing problems.
By the time he became Chief Analytics Officer at Grupo San Cristóbal and then Chief Data Officer at Reba, Ruyu had done something rare in the AI field: he had actually managed data at scale inside old, complex organizations with legacy systems, politics, and entropy. He knew exactly which data engineering problems were unsolved - because he had spent a decade and a half trying to solve them by hand.
In 2021, before the term "AI agent" became a product category, Ruyu was already building them in Latin America. By 2022, that work had crystallized into a company.
Every mid-size company has the same problem: more data than its team can handle, a data engineering backlog measured in months, and SQL queries that require scheduling a meeting with someone who gets pulled into other meetings. Teramot's answer is to remove that bottleneck entirely.
The platform connects to your existing databases and automatically writes production-ready SQL. It builds ETL pipelines without requiring anyone to code them. And it integrates with MCP (Model Context Protocol), which means a business analyst can open ChatGPT or Claude and ask a question about live company data - and get an answer, without involving an engineer.
Teramot builds and maintains data pipelines automatically. No code, no tickets, no waiting for the data team to get back to you.
Connect your database to ChatGPT or Claude. Ask questions in plain English. Get answers from your actual data, in real time.
Enterprise-grade security built in. Teramot is SOC 2 certified - meaning it clears the compliance bar for deployment inside serious organizations.
Ruyu describes Teramot as "the layer that connects all these agents with any company." That framing is telling - he's not building a product, he's building infrastructure. The kind of infrastructure that matters more the more AI agents proliferate.
The Big Idea
Most companies approach AI as a feature - something you bolt onto an existing product. Ruyu thinks about it as a team replacement. Not a replacement for people, but a replacement for the coordination overhead that makes data work so slow. The meetings to prioritize the backlog. The back-and-forth on requirements. The three-day wait for a dashboard refresh.
Teramot absorbs that overhead. The "Artificial Data Team" isn't a gimmick - it's a claim that you can get the output of a four-person data team without the logistics of managing one.
The platform is designed to scale with companies that are at very different points of data maturity - from those with no data strategy at all to those swimming in it. That breadth is strategic: enterprise data engineering is a $50B+ market, but most of the deals have historically gone to companies large enough to build internal teams.
Teramot's target is everyone who couldn't afford the old approach - which is most companies, most of the time.
Ruyu's public writing gives away his particular brand of intellectual curiosity. On Medium, he has explored the legacy of Edmond Halley - not the comet, but the man who funded Newton's Principia out of his own pocket and never saw recognition for it. The essay, titled "The Purpose of Halley," is a meditation on what it means to enable someone else's greatness. It is not the writing of a hustle-culture founder. It reads like someone who thinks in centuries.
He has written about "The Trust Gap" in digital transformation - the invisible distance between what technology promises and what organizations actually believe it can deliver. And he wrote "5 Ways to Ensure Your Data Science Initiative Will Fail," which reads like a field guide assembled from watching smart companies make the same predictable mistakes, over and over, across three different industries.
His LinkedIn tag line is the tersest possible version of his whole story: "Doing AI since 2005." No credentials listed. No awards. Just the date, offered as a challenge.
"Doing AI since 2005."
- Bruno Ruyu's LinkedIn headline. Four words. Twenty years of patience.Teramot launched with a lean but deliberate crew. Gabriel Puertas, Head of Product and Operations, brings an engineering background from the University of Sheffield and deep operational experience from Argentina. Sacha Buengueroff handles Sr. Python Backend Development. And in 2024, Jon Castor joined as Independent Director - 25+ years building venture-backed startups and Fortune 500 divisions, including a stint as Chairman of Omneon and CEO of TeraLogic. Castor's involvement signals that Teramot has moved from "scrappy Argentine startup in San Francisco" to something with institutional velocity.
Investors include Natan Ventures, MANA Tech, Cites Startups, and a handful of accelerators including Newchip and Startup Ole. With $600K+ raised and 26 people on the team, Teramot is past proof-of-concept and into the growth-before-Series-A stage where execution discipline matters most - which, given Ruyu's background, may be his actual home court.
Education & Training
Built and launched the first AI-native data engineering platform that automates SQL, ETL, and database querying through natural language.
Among the first in Latin America to build and ship production AI agents, in 2021 - before the category had a name in VC circles.
Led Teramot through enterprise security certification - a prerequisite for serious enterprise sales that most early-stage AI startups skip.
Rare enterprise depth: modeling, uncertainty analysis, cloud-native architecture, and AI leadership at YPF, Reba, and Grupo San Cristóbal.
Completed Stanford GSB Executive Program 2023-2024, adding strategic business leadership credentials to his technical foundation.
Shipped Teramot's MCP integration connecting databases to ChatGPT and Claude - positioning the company at the center of the enterprise AI agent wave.
Emprender en inteligencia artificial desde Argentina, desafios, mitos y futuro
Spanish - YouTube
Crear una STARTUP de IA: La REALIDAD - Bruno Ruyu / Teramot
Spanish - YouTubeHe trained his first neural network in 2005 - the same year YouTube launched and the same year Deep Learning was a niche academic field with almost no commercial relevance.
Instituto Balseiro in Bariloche, Argentina accepts roughly 50-80 students per year nationwide. It is considered one of the finest physics and engineering schools in Latin America.
He holds both a physics degree and a Master's in Finance - making him equally fluent in backpropagation and balance sheets, which turns out to be useful when you're selling enterprise AI.
Teramot's AI was named "The Sentinel" - a reference that feels like equal parts Isaac Asimov and SOC 2 compliance framework.
His Medium essay about Edmond Halley - the man who funded Newton but never got the credit - reads like a philosophical statement about the kind of builder he wants to be.
Teramot's MCP integration means any employee at a client company can now ask ChatGPT or Claude a question about live company data - and get an actual answer, without a SQL ticket.
Topics & Keywords