© All rights reserved. Powered by Techronicler
In Conversation with
Jason Hishmeh
A Techronicler interview with Jason Hishmeh, CTO/CISO, Varyence and Equet.
Techronicler: Thank you for joining us, Jason! Well, everyone has an origin story! What was the first piece of technology you ever broke, built, or fell in love with?
Jason Hishmeh:
For me it started when I got my first computer, a Commodore 64, back in the 80s when I was very young. I spent hours trying to play with it, hack it, and take it apart. What hooked me was not the hardware, it was the realization that one person could write software and thousands of people could use it. That still excites me today.
Techronicler: What is the one problem or project that is taking up 80% of your brain space this month?
Jason Hishmeh:
Agentic AI governance. Teams are handing real work to autonomous agents faster than they are putting guardrails around them. The question consuming me is not “can the agent do the task,” it is “what happens when it does the task wrong at scale, and who is accountable.” So we are building the boundaries: what agents can touch, what requires a human in the loop, how decisions get logged and audited. The upside is huge, but an unsupervised agent with production access is a security incident waiting to happen.
Techronicler: How do you define the role of a manager when 40% of your team’s ‘output’ is generated by autonomous AI agents?
Jason Hishmeh:
The manager’s job shifts from directing work to designing the system that produces it. When agents generate a large share of the output, your leverage is in the boundaries, the review gates, and the judgment calls machines cannot make. I have started budgeting for AI agents the way I budget for people: they have scopes, permissions, and a cost, and they need oversight. What you cannot delegate is accountability. A good manager now spends less time assigning tasks and more time deciding what “good” looks like, then holding both humans and agents to it.
Techronicler: Can you describe a time you had to intervene in an AI-driven decision that felt technically correct but ethically or culturally misaligned?
Jason Hishmeh:
Yes. We had fine-tuned an AI model for a US healthcare startup. It was optimizing for the company’s profitability exactly as the metric asked, and that pushed it toward choices that cut against the reason the company existed: helping people live healthier lives in their later years. The model had no idea it was about to erode trust to win a number, so I overrode its thinking and refined it further. The lesson was not that the AI got it wrong, it was that it did not understand what real success looked like, or what the business was actually for. Ethical AI is critical for public confidence as agent autonomy goes mainstream. A human has to own the decision when a system optimizes for the letter of the goal and misses the intent. Auditable decisions and a clear human owner are not optional.
Techronicler: Tell us about a time you had to make a deeply unpopular technical decision that turned out to be the right call. How did you handle the pushback?
Jason Hishmeh:
I have had to slow projects down and change a project’s technical direction mid-stream because of bad early architecture decisions. Both were unpopular in the moment. The one I would point to: choosing to slow down and build security and compliance in from the start on a Healthcare tech project, on a platform where everyone wanted to get to market quickly. It meant that when the enterprise healthcare deals came through the sales team, we could actually close them and meet their security and compliance requirements. I had to be direct with every stakeholder about the tradeoff and the risks, and take public ownership of the call, so everyone knew exactly whose decision it was and who was accountable for it. In the end we closed some large deals, and people realized we would have lost them otherwise.
Techronicler: What is a piece of ‘common wisdom’ in the tech industry that you completely disagree with?
Jason Hishmeh:
The old wisdom is that you have to raise big to build something huge. I do not buy it. I believe in bootstrapping a startup as far as you can on the least outside capital possible, because that is how you earn the most return on your own time and effort. I would rather see a founder build a $10 to $50 million business and keep more equity. Too often a founder spends years building a company, raises a lot of VC, announces a huge exit, and walks away with almost nothing. Do not be that founder.
Techronicler: As a leader, how do you architect workflows to reduce the ‘mental load’ and cognitive burnout of your team in a 24/7 digital environment?
Jason Hishmeh:
A lot of the burnout I see comes from poor leadership, bad processes, and unnecessary politics, not from the work itself. My advice to CEOs is simple: appreciate your people, build real relationships with them, get to know their concerns, and design your processes so that people can succeed. Fix that, and most of the burnout takes care of itself.
Techronicler: How do you evaluate ‘potential over polish’ during the hiring process to ensure a more equitable team?
Jason Hishmeh:
I believe potential comes more from a person’s character and work ethic than from the skills they already have. That is getting harder to see in a world of AI HR screening. Actually getting to know the person in front of you is a skill fewer and fewer people have. Stop hiring logos on a resume. Hire for passion, character, and work ethic, and you will build a stronger, better team.
Jason Hishmeh is an entrepreneur, investor, and technology leader with 25+ years building and securing technology for companies ranging from early-stage startups to the Fortune 500. He is co-founder, partner, and CTO/CISO at Varyence.com and Equet.com. His work centers on putting AI to work where it moves real business outcomes, with security and compliance built in from day one. He writes and speaks regularly on AI adoption, cybersecurity, and pragmatic technology leadership.