A deal platform carrying $200M+ of daily flow

Context

A deal-processing platform for one of Australia's largest investment banks, carrying $200M+ of daily deal flow under strict security and compliance requirements.

The engineering problem and the organisational problem were the same problem: teams spread across four time zones, requirements that moved constantly, and a system where a defect is a financial event rather than a bug report.

I arrived as team lead over three developers. One of them was carrying the project; the other two were producing the appearance of work. That is a staffing problem disguised as a delivery problem, and no process fixes it.

Scale

What I decided and built

• ›Changed out the two engineers who were not delivering instead of building process around them, and started hiring replacements immediately — while writing production code myself to keep delivery moving.
• ›Stayed a contributor throughout the growth: top contributor by commits and merged PRs while leading.
• ›Built the client relationship until I was running the project as PM as well as leading engineering — which is what made the hiring plan possible.
• ›Hired to 30 engineers and QA across six teams, each with its own lead, QA and developers, so every team could own delivery end to end rather than passing work sideways.
• ›Owned Scrum, environments and CI, so that teams in four time zones could integrate without blocking each other.
• ›Acted as the bridge between four countries under constantly shifting requirements — the coordination layer was the deliverable as much as the code.
• ›Delegated deliberately as the team grew — technical interviews first, to the strongest engineers — writing less code myself each month until several teams were doing hands-on work instead of me.
• ›Prepared the handover over several months rather than at the moment of leaving, and handed the project to newly hired managers at a major release. The structure had to work without me, and it did.

Outcome

• ◆$200M+ of daily deal flow processed under strict security and compliance.
• ◆Grew a team 5× in six months, and hired to 30 engineers and QA overall — shipping continuously to a client who stayed close to the project throughout.
• ◆Handed over cleanly at a major release, to managers hired and onboarded for it — a transition planned over months, not an exit.

What I'd do differently

• ›I hired for engineering ability and for how someone reads in an interview room. One of those hires sent a message to the client directly, written in anger. Overnight the client's minimum test-coverage requirement went from 85% to 100%.
• ›It was recoverable and I recovered it, but the cost was real and it came from one message. Interviewing well is not evidence of composure with a customer in the thread, and I would screen for the second thing explicitly — and be deliberate about who talks to the client directly, and when.