The hardest part of security leadership isn’t building better detection systems; it’s staying connected enough to the daily work that you can still recognize when something doesn’t make sense. William Glazier, Director of Engineering – Threat Research & Machine Learning at Cequence Security, refuses to detach from the operational reality his team faces. He’s still in the on-call rotation, still debugging production issues, still analyzing customer data to understand how real investigations unfold.

William discusses why mental agility matters more than formal credentials when building security teams and how protecting space for people to follow their curiosity creates culture that outlasts any individual leader. He shares his framework for building trust through selective transparency about decision-making, why writing everything down during incidents prevents repeating the same mistakes, and how to recognize when you’re letting emotions rather than data drive your responses.

Stories We’re Telling Today: 

• Building SOCs from scratch and scaling them as threat landscapes evolve from simple bot attacks to sophisticated AI agents
• The practice of comprehensive documentation during security incidents, turning false positives into trust-building opportunities 
• Screening for mental agility over credentials, by testing candidates’ willingness to adjust when given with contradictory evidence
• Creating team cultures where people teach each other across disciplines, enabling understanding across security contexts
• Maintaining operational connection as a leader by staying in on-call rotations and debugging alongside team members
• The strategic use of AI tools for documentation and transcription while avoiding the trap of outsourcing critical thinking
• Why customer obsession requires vendors to use their own products exactly as customers do, including friction points, to maintain perspective

Too busy; didn’t listen: 

• Stay operationally connected with actions like remaining in on-call rotations and debugging alongside engineers to maintain and build trust.
• In hiring, screen for people who change their minds when presented with contradictory evidence rather than forcing conclusions to fit initial assumptions. This shows mental agility.
• Comprehensive documentation during security incidents transforms potential failures into trust-building opportunities.
• The biggest risk of AI tools isn’t replacement but outsourcing critical thinking before you’ve struggled through enough problems to recognize when AI-generated answers don’t make sense.

Skip to the Highlight of the episode: 

[33:30-33:56] “If you have a relationship with someone that’s generally based on not being afraid to trade ideas and “Actually I disagree with you here. I disagree with this. What about this? What about that?” And then that stops. That’s another red flag. And how do you build it? It’s relying on you, again, having some North star that you’ve got to keep. People have to know some goal they’re working towards, some broad vision.” 40:51-51:22

Speaker

William Glazier, Director of Engineering – Threat Research & Machine Learning, Cequence Security

William Glazier has built, scaled, and managed security operations while developing machine learning models for bot detection. His background includes everything from early threat intelligence work on residential proxy services to current challenges distinguishing good AI agents from malicious ones.

Listen to more episodes: 

Apple 

Spotify 

YouTube

Website