Ascendion's CTO: Design thinking, not coding speed, is engineering's future

Ascendion CTO Wesley Pullin has spent more than 34 years in software development, working across large software companies and most recently at CloudBees, where the Jenkins ecosystem touched upward of 55 million developers. When Ascendion co-founder Arun Varadarajan recruited him, it was not a familiar pitch. What closed the deal, Pullin said, was a single phrase he had never encountered in three decades of the industry: assuring client outcomes. That framing, not selling licenses or chasing ARR, became the lens through which he now thinks about engineering's next chapter.

Ascendion itself is a relatively young company, conceived during the COVID period and built into an organization approaching 10,000 employees and nearly a billion dollars in revenue. Its trajectory is less a linear startup story and more a deliberate disruption of how enterprises build and operate software. The company's platform, AAVA, is designed to help organizations re-examine legacy systems and the accumulated technical debt that makes those systems slow and expensive to change.

Design before code: recovering a lost discipline

Varadarajan is direct about where the industry went wrong. In his view, the shift toward agile methodologies gradually eroded the time engineers spent on design and architecture, pushing teams to start writing code before the problem was fully understood. The consequences are visible in today's enterprise systems: costly to change, difficult to extend, and poorly suited to new market demands.

The lack of focus on design and architecture is what has yielded us where we are today, and that's why today's systems are so expensive to change, expensive to rewire, expensive to meet market demands. — Arun Varadarajan, Co-Founder, Ascendion

Pullin reinforced the point with a concrete analogy. Adding AI copilots and coding agents to an existing workflow speeds up one segment of the software development lifecycle, but the remaining steps become bottlenecks. The net return on that investment ends up marginal, because only one constraint was removed while others intensified. The fix, both men argue, is not more tooling layered on top of a broken process. It is a return to disciplined design thinking: user stories, epics, high-level and low-level design, naming conventions, and security considerations, all before a line of code is written.

When we just do a vibe code and go straight to coding, some of that can be eliminated. And the danger is you eliminate that, and that's everything you need to refactor, to fix, to stop prompt injection, to stop people from hacking it. — Wesley Pullin, CTO, Ascendion

A new operating model for enterprise engineering

Ascendion's response to this structural problem is what Varadarajan calls an engineering-to-the-power-of-AI method. It rests on two changes to how enterprises operate. First, processes must be redesigned from the ground up. The workflows built for human-centric execution need to become human-powered, meaning people are freed to do higher-order thinking while systems handle repetitive construction tasks. Second, the people model has to change. The company has developed new job families, career lattices, and competency frameworks, moving away from purely skill-based definitions of an engineer toward roles defined by problem-solving orientation and solution design.

That shift in how engineering talent is defined also carries implications for who gets to participate. Varadarajan described asking a room of sales professionals how many considered themselves engineers, and then challenging the assumption that an engineering degree is what makes someone an engineer. His example: his best engineer at IBM was a CPA. The goal Ascendion has set for itself is to democratize engineering, making it accessible to people who have deep domain knowledge but lack traditional coding credentials. In that framing, the question is not whether someone can write code. It is whether they can define the problem clearly enough that the right solution gets built. If the platform handles the construction, the human's value lies entirely in the quality of the thinking that precedes it.