Co-Founder & CEO of MindsDB. Visiting Research Scholar at UC Berkeley. The engineer who decided that if machine learning was too hard for 95% of developers, he'd build a layer that made it feel like a database query.
In 2008, Jorge Torres was working at Couchsurfing - the travel community startup setting up couch-sharing across the world - when he walked uninvited into a UC Berkeley classroom. No student ID. No enrollment. He asked Professor John DeNero if he could sit in on CS188, Introduction to Artificial Intelligence. The professor said yes. Torres calls it "the best thing that happened to me."
What he saw in that room - computers learning to solve problems - became the engine of everything since. By 2018, he had co-founded MindsDB with Adam Carrigan, named after the all-knowing Minds in Iain M. Banks' Culture science fiction series. The idea: if intelligence could be woven into a machine, why couldn't it be woven directly into the databases where all enterprise data already lives?
Torres's bet was precise. Of roughly 30 million software developers worldwide, fewer than 5% are proficient AI and ML engineers. The other 95% know SQL. So MindsDB built a system where you query a machine learning model the same way you query a table - with SELECT statements. The abstraction was radical. The adoption was decisive.
MindsDB graduated from Y Combinator's Winter 2020 batch, then raised $16.5M from Benchmark Capital in early 2023, followed by a strategic investment from NVIDIA's NVentures arm later that year, pushing total funding past $46.5M. Today the platform has over 150,000 open-source deployments and connects AI agents to more than 200 data sources through a single Model Context Protocol (MCP) server.
Before MindsDB, Torres's career read like a tour of the internet's scaling era. He was the first full-time engineer at Couchsurfing, helping take it from a few thousand to several million users. He moved infrastructure to CDN at Skillshare.com. He built computer vision systems for out-of-home advertising at Real Life Analytics. He spent time at CareJourney working alongside Aneesh Chopra - the first CTO in US government history - building data systems that processed billions of patient records.
Each stop added a layer: distributed systems, video infrastructure, computer vision, healthcare data at scale. MindsDB is what happens when all of those layers compound. Torres is also a Visiting Research Scholar at UC Berkeley, researching machine learning automation and explainability - the same university where, as an unofficial auditor, he first understood that computers could learn.
"Machine learning has this incredible potential to solve some of the world's biggest challenges - climate change, health, education. But only if the technology is well understood and accessible to every person and every organization."
Jorge Torres - Co-Founder & CEO, MindsDB
Torres built MindsDB around a single observation: enterprise data already lives in databases, and developers already know how to query databases. So instead of asking businesses to become AI shops, MindsDB brings the AI to where the data already is.
The platform acts as a federated query engine - an intelligent middleware that can connect to over 200 databases, applications, and data stores simultaneously. You write SQL. MindsDB talks to the models. Your existing tools stay unchanged.
In 2025, MindsDB became one of the first platforms to support Anthropic's Model Context Protocol (MCP), letting a single MCP server replace dozens of one-off integrations. One server. 200+ sources. Torres describes the problem it solves bluntly: "Organizations implementing AI face a fundamental challenge - connecting models to the right data, at the right time, with the right security controls."
"The longer a company exists, the more data it collects. But sometimes having too much data is the same as having no data at all."
Jorge Torres
"Make your obsession your profession, and you'll never have to work a day in your life."
"If you try to engineer a complex system end-to-end, you'll fail. You need to start simple, iterate, and let the community help you refine it."
"I believe that there is enormous power in data. The more a company has, the more they're able to propel their businesses forward."
"We're basically augmenting the decision-making process to make it easier and more efficient."
"Organizations implementing AI face a fundamental challenge: connecting models to the right data, at the right time, with the right security controls."
"I was amazed that computers could learn to solve a problem." - on attending his first AI lecture at Berkeley in 2008
In 2008, Torres was living in Berkeley and working at Couchsurfing. He walked into the UC Berkeley CS188 AI class without being enrolled and asked Professor John DeNero if he could sit in. DeNero said yes. Torres describes that moment as "the best thing that happened to me." He walked in a software engineer. He walked out on the path to becoming an AI founder. Twelve years later, the same university would make him a Visiting Research Scholar.
MindsDB is named after a science fiction concept. Iain M. Banks's Culture series features "Minds" - vast, self-evolving artificial intelligences that manage entire civilizations while remaining aligned with human values. When Torres and co-founder Adam Carrigan named their company, they weren't being poetic for investors. They meant it literally: build systems that think, predict, and stay aligned with the people using them. The sci-fi name is a design specification.
Torres repeatedly cites one statistic: fewer than 5% of the world's 30 million software developers are AI or ML engineers. Every enterprise AI pitch assumes that businesses will hire the 5%. MindsDB was built for the 95% who already know SQL and don't want to hire a data science team to get a prediction from their own data. The insight sounds simple. Executing it - building a query engine that abstracts ML as a database layer - took six years and $77M.
Before founding MindsDB, Torres worked at CareJourney alongside Aneesh Chopra - the first CTO the US government ever appointed. Together, they built data systems that processed billions of patient records to optimize healthcare cost and quality. It was not a glamorous startup pivot. It was deep, messy, high-stakes data engineering at a scale where errors have human consequences. Torres took that seriousness about data quality and governance directly into MindsDB's enterprise product.
"If you try to engineer a complex system end-to-end, you'll fail. You need to start simple, iterate, and let the community help you refine it."
Jorge Torres - on building open-source AI infrastructure
