Startup Speed Inside a Century-Old Bank
After two years at Kabbage helping build what became a fintech unicorn, I took a deliberate turn: could the practices that made a startup move in minutes survive inside a commercial bank that's been around since before the airplane? At SunTrust — now Truist — I spent five years finding out.
The Problem: Legacy at Scale
Commercial banking ran on systems and habits shaped by decades of waterfall delivery: quarterly releases, siloed data sources, brittle integrations, and an uptime bar set by customers who bank from their phones at midnight. The technology wasn't failing because people were careless; it was failing because the organization had never been asked to move differently.
Bringing Fintech Discipline to Banking
My role as technical team lead was to import what I'd learned at Kabbage — not as a lecture, but as working software and working habits:
- Agile transformation: Introduced sprint cadence, backlog ownership, and demo culture to teams accustomed to big-bang releases. Developers started seeing their work ship in weeks, not quarters.
- Automated testing and deployment safety: We invested heavily in test coverage and deployment gates — moving risk left so integration failures surfaced before production, not after a pager went off. Continuous deployment was never on the table in a regulated banking environment, but we got to a state where deploying was a calm, practiced activity rather than a stressful event.
- Service-oriented decomposition: Broke apart large multipurpose applications into focused, single-responsibility services with clear interfaces. True microservices require data isolation, and that wasn't achievable when everything drew from the same centralized mainframe database — but separating the application layer was still a meaningful step toward maintainability and team autonomy.
Earning Trust, Changing Habits
The hardest part wasn't the tooling. It was earning the trust of engineers and product owners who had seen "transformation initiatives" come and go. I focused on small, visible wins: one team shipping weekly, one service reused by a second team, one incident post-mortem that actually changed a process. Momentum built on visible wins beats fear-driven mandates every time.
Outcomes Over Five Years
The work wasn't a single product launch — it was a sustained shift in how software got built:
- Release frequency increased from quarterly cycles to on-demand deployments on key product lines
- Production incident rates dropped as automated testing and deployment gates became the norm, not the exception
- Developer onboarding accelerated as well-defined service interfaces replaced tribal knowledge and one-off workarounds
- Teams grew in both headcount and capability, with several engineers I mentored moving into lead roles themselves
The Bridge Forward
SunTrust taught me that organizational change and architectural change are the same problem wearing different hats. You can't decompose services into a waterfall culture and expect magic. But you can show a 100-year-old institution what "good" looks like — if you're patient, if you ship proof, and if you remember that the developers you're mentoring didn't choose the legacy; they inherited it.
That combination — fintech speed, enterprise gravity, and five years of sustained practice — is what prepared me for director-level roles where the mandate is transformation at even larger scale.
