Skip to content
Work

Meta · Dublin

Operating between developers and everything that had to move for them

Role
App Review & Developer Support Specialist, later Project Manager
Period
2020–2022
Location
Dublin, Ireland

The problem

Developer escalations arrived faster than any person could work them, across six platforms with different rules, and every one of them needed a decision from a team that did not report to me.

Context

  • The surface was Facebook, Instagram, Messenger, Oculus, Instant Games and Camera Effects. Each had its own policy nuances and its own engineering owners.
  • Resolution almost never sat with me. It sat with Engineering, Product, Legal, Partnerships or Policy, each with their own priorities and timelines.
  • I was also the point of contact for Instagram specifically, handling Instagram developer issues and advocating internally for its APIs.
  • Volume was not the interesting part of the problem. Coordination was.

How I approached it

  1. 01

    Triage on impact and urgency, not on volume

    I built a four-quadrant matrix: impact against urgency. Platform-wide issues with partners or revenue exposed were handled immediately. Individual developer blockers, strategic process issues and documentation requests each got a different path. The point was to stop reacting to whoever escalated loudest and start deciding where attention was worth spending.

  2. 02

    Route with enough context to act on immediately

    One person cannot solve everything, so the real skill is knowing who can and giving them everything they need in one pass. I built a routing model. Engineering bugs went to Developer Support Engineers with reproduction steps, policy violations to Legal with case history, partnership issues to Partnerships with relationship context, product questions to PMs with user-impact data. The gain was not knowing where to send things. It was eliminating the back-and-forth.

  3. 03

    Read the data for where problems are created

    At that volume, patterns are visible if you look. I analysed over 2M records to find which platforms generated the most escalations, where cases stalled, and which issues recurred. That moves you from resolving problems to removing the conditions that produce them.

  4. 04

    Automate the predictable, keep judgement human

    Initial triage classification, data gathering, routing for simple cases, status updates, case summaries and metrics reporting were automated. Anything requiring nuance, trade-offs or cross-functional alignment stayed with a person. If a decision follows clear rules and needs no context, it should not need a human.

What I did

  • Managed 200+ high-priority developer escalations monthly across six Meta platforms
  • Handled developer and product support directly: API integration problems, platform behaviour, implementation blockers
  • Served as the Instagram point of contact for developer issues and API advocacy
  • Led automation projects and compliance enforcement work
  • Coordinated delivery across Engineering, Product, Legal, Partnerships and Policy

Result

95%

Resolution rate

36%

Efficiency improvement

50%

Reduction in malicious app cases

20%

Less manual work

30%

Faster cross-functional delivery

40%

Improvement in developer satisfaction

What I took from it

  • At scale, individual effort is not a strategy. The systems either hold or they do not.
  • The hardest part of platform operations is not technical. It is getting teams with different incentives to move on the same thing.
  • A well-formed escalation is a piece of engineering. Most of the delay in cross-functional work is missing context, not missing willingness.
  • Patterns in operational data tell you where a problem is manufactured. The place it surfaces is rarely the place to fix it.

Related

Other work

If any of this is the shape of a problem you have, I'd like to hear about it.

Get in touch