Andrea Coifman explains how MCP helped Nasdaq move from asking questions to trusted enterprise agents that use real tools and workflows, and why the people side was harder than the protocol.
About Andrea Coifman
A passionate technology leader with a background in software engineering, data science, and AI, Andrea Coifman leads the AI Platform as Director of Software Engineering at Nasdaq in Vilnius. With a career spanning roles at Devoteam, Qlik, and Jabra, she has focused on creating impactful AI-powered solutions, from auto-clustering algorithms to personalized design tools, with her work published at several conferences. Driven by a belief in technology’s power to solve real-world problems, Andrea holds a Master’s in Engineering from Aalborg University and an engineering degree from the Technical University of Madrid. She is also committed to fostering the next generation of tech talent and has mentored for multiple Women in Tech organizations.
What motivated you to join the AI Builders Global Conference?
Very much so. What drew me in is the focus on builders. A lot of AI conversations stay at the level of vision and hype, but the most useful lessons come from people who have actually shipped something, hit the limits, and had to rethink their approach. That’s the kind of exchange I want to be part of: sharing what worked for us, what didn’t, and learning from others who are solving the same problems in different contexts.
What shaped your journey through technology and AI?
I trained as an engineer at the Technical University of Madrid and then did a Master’s in Engineering at Aalborg University in Denmark. That’s where I fell for data science: the idea that you could find patterns no one had told you to look for. From there I worked across very different companies, including Devoteam and Qlik. At Qlik I helped bring auto-clustering algorithms directly into the analytics product, and other projects took me into AI-powered personalized design tools. Some of that work ended up published at conferences. The lesson that stayed with me is that a model only matters once it’s in people’s hands. Then ChatGPT launched, and almost overnight leadership stopped asking whether AI was real and started asking how fast we could use it. My work shifted from building models to building systems around them. What keeps me here is that AI is both a technical challenge and a human one. Building agents that work reliably at enterprise scale is hard engineering. Getting an organization to trust them and change how it works is just as hard, and I find the two problems most interesting together.
Why is Nasdaq’s agentic journey particularly important to you?
Because it’s the moment I’ve seen AI become genuinely useful inside a large organization. For a long time, AI at work meant asking questions and getting answers. That’s helpful, but it doesn’t really change how work gets done. MCP was the turning point for us. Agents could finally act: use real tools, follow real workflows, and do it in a way a regulated company can trust. What makes it personal is that the hardest part was never the protocol. It was the people side: getting teams to stop building in silos, trust shared foundations, and change how they work.
Who should attend the AI Builders Global Conference?
Anyone responsible for turning AI from a pilot into something people use every day: platform and software engineers, architects, product leaders, and engineering managers. I’d especially recommend it to people in regulated or complex industries, where security and governance can’t be an afterthought. You’ll leave with practical patterns, honest lessons about what breaks, and a network of people working through the same challenges.