# Team recovery: delivering the projects that had already failed

> Joining as an engineering manager amid missed deadlines and lost confidence — and leaving the team delivering beyond its own expectations.

Durbin Labs · Aug 2021 — Dec 2021

## Context

I joined Durbin Labs as an engineering manager during a difficult stretch: several customer projects had already missed their deadlines, customer relationships were at risk, and the engineering team had become demoralized — it had lost confidence in its own ability to deliver.

## The problem

The visible problem was late projects. The actual problem was a team that had stopped believing its estimates, priorities that shifted without a clear owner, and technical blockers that nobody felt empowered to remove. Adding pressure would have made all of it worse.

## What I did

- Took ownership of the delayed customer projects and set practical delivery priorities — what ships, in what order, and why.
- Removed technical blockers personally while staying hands-on: architecture, backend implementation, code reviews, AWS infrastructure, production troubleshooting.
- Mentored engineers and deliberately handed them ownership, so recovery did not depend on me heroically holding every thread.
- Rebuilt delivery processes around what the team could actually commit to.

## What changed

Leadership range, and the arc of the recovery:

## The result

Multiple previously missed projects were delivered within a relatively short period, customer trust was restored, and the team delivered beyond its own initial perception of its capability — which was the outcome that mattered most for everything that came after.

- **Multiple** — missed projects delivered
- **Hands-on** — management: architecture, code, production
- **Trust** — restored with customers and within the team

## What I would do differently

I would formalize the sequencing and the stakeholder communication cadence much sooner. Recovery only started working once priorities were explicit and status was honest; doing that in week one instead of discovering it through iteration would have removed weeks of thrash.