Architect, Engineering
Dublin, Ireland
Telmo Felix is an Engineering Architect at Dun & Bradstreet, where he has helped shape the company’s data infrastructure through innovative development and early adoption of emerging technologies. He is currently leading research efforts in GenAI, where he helped develop ChatD&B — a conversational agent designed to interact with structured data and enhance how businesses access insights.
In this Q&A, Telmo reflects on his path from open-source development to leading data innovation at Dun & Bradstreet, including his work on patented platforms and the GenAI-powered ChatD&B assistant.
I’m originally from a small village in Portugal. I went to Lisbon to study computer science. There, I got involved in the open-source community and was a founding member of the Portuguese Linux community. We had meetings where people like Richard Stallman and Alan Cox — two of the most well-known developers of Linux — attended. So I was always interested in the creative aspects of open source.
I started working while in college in Portugal. This was during the internet boom, and I did what is today called full stack development — working on everything from the back-end to the front-end on websites and in e-commerce. Around 2005, I brought mobile development to my then-employer which was extremely new at the time. Around that time, I was finding opportunities limited in Portugal, so I joined D&B and moved to Ireland about 15 years ago.
Initially I was working mostly on web development here on DNBi and similar applications. But I was always more of a data and back-end developer, so that’s when I got onto a team that experimented and created a proof of concept to gather all our data in one place, ensure its consistency, and render it product-ready. This proof of concept eventually turned into Prime and resulted in 2 granted patent families. We were very much ahead of time on things like the data lake concept, where you bring all data sources into one place and choose from there.
After Prime was developed, I left D&B for a while and explored other companies for about 6–7 years. During that time, I was involved in different projects from data processing to healthcare to hardware. The opportunity arose to return to D&B in a research and development role, which is where I work now. As part of our research two years ago, we explored what the D&B take on GenAI would be, and that exploration resulted in ChatD&B: the agent that can interact with D&B data. Again, we were very early adopters of that when we started getting our agents to work and to interact with D&B data. I think that's one of the things that we innovated: How can you interact in a conversational way with structured data? We’re now patent-pending on that concept.
1Aggregating high volumes of temporal data from multiple overlapping sources (Link: US00000010942911B220210309)
2System and method for preserving interdependent corporate data consistency in a globally distributed environment (Link: US00000010891283B220210112)