Back to home

Overflow · 2022–2023

Bringing Process to Chaos Without Killing Momentum

An 8-person team with no delivery structure, fragile infrastructure, and compliance gaps. The challenge: fix all three without breaking what was already working.

Role

Director of Engineering

Team

6 engineers + 1 designer

Stack

AWS · Scrum · SOC 2

Scope

Process · Infra · Compliance

The situation

Overflow had real talent and a product with genuine traction — but delivery was unpredictable. Project tracking lived in a Notion doc nobody fully trusted. Leadership couldn't reliably forecast when things would ship, and engineers weren't always sure what they were supposed to be working on next.

The infrastructure also had inherited risk: a setup that had grown organically, with fragile failover for critical services like email and SMS, and costs that hadn't been scrutinized since early days. SOC 2 compliance was a requirement for enterprise sales, but the work hadn't started.

Fixing delivery visibility

I introduced Scrum not as a bureaucratic mandate but as a shared communication tool. The goal was visibility — for the team, for product, and for leadership — without adding overhead that would slow people down. I worked with the team to define what "done" actually meant, establish a backlog that everyone trusted, and create a rhythm that made work predictable without making it rigid.

Within two sprints we had reliable velocity data and a planning process that gave leadership genuine confidence in forecasts. Engineers had clearer priorities and less context-switching. The change wasn't about process for its own sake — it was about removing the ambient anxiety that comes from not knowing what you're supposed to be doing.

Infrastructure & compliance

In parallel, I led a full AWS infrastructure migration, consolidating two platforms into one — saving $2,500/month in hosting costs, eliminating 20 hours/week of dual-platform maintenance, and improving email/SMS failover reliability by over 90%. These weren't just technical wins — they removed operational anxiety that had been quietly taxing the whole team.

I also partnered with risk and compliance to achieve SOC 2 Type 2 certification, developing all engineering policies and controls from scratch. This required building a clear picture of our data flows, access controls, and operational practices — and then formalizing them in a way that would hold up to audit. The certification opened enterprise sales conversations that hadn't been possible before.

Process only works if engineers trust it. I focused on making the system serve the team — not the other way around.