Integration Tech Lead (Ad Platform Integrations, GCP)

We are looking for an Integration Team Lead. This is a technical management role: you lead a team of 24 engineers, own the integration architecture and roadmap, and stay hands-on in the code. Every number our clients see in a dashboard starts as an API call one of your pipelines made, so reliability and data correctness land on your desk first. You report directly to the CTO and prioritise the roadmap together with the product team. Fully remote, EU timezones.

Requirements
General
2+ years leading engineers as a tech lead, team lead or hands-on engineering manager: code review, sprint planning, mentoring
5+ years in backend or data engineering, with production systems you supported yourself
Experience integrating third-party APIs at scale: rate limits, OAuth flows, webhooks, async reports
Comfortable in a fully remote team with hours overlapping EU timezones
English B2 or above. Most of our communication is written and async

Technical
Strong Python 3.11: production-grade code with error handling, structured logging and tests
Data pipeline patterns in practice: idempotency, backfill, incremental sync, schema evolution
GCP at the service-design level: Cloud Run services and jobs, Cloud Functions, Cloud Scheduler, Cloud Tasks, Cloud SQL, GCS, Secret Manager
SQL well beyond the basics, and comfort putting real logic into it. A large part of this pipeline lives in BigQuery scheduled queries and versioned BigQuery migrations rather than in Python
Writing FRDs and keeping technical documentation current: integration contracts, data models, runbooks. We do not have this discipline today, and establishing it is a core part of the role rather than a bonus
Monitoring and alerting you set up yourself: Grafana, logs, retry and backfill strategies. Today every alert path ends in Slack
Docker, and CI/CD you will own. These services currently deploy from hand-run scripts with no pipeline and no tests in CI. Building that out is part of the job, not something you inherit finished

Soft
You lead in practice: you split work for other people, review their code, unblock them, and say no when that is the right answer
Systems thinking: you see the full chain from API ingestion to the end-user dashboard
When a number looks wrong, you trace it back to the source before you change anything
Task management: your roadmap and your team's board reflect what is actually happening, and you raise risks while there is still time to act on them
Pragmatic and transparent, including when something breaks: you say what happened and why, early

AI baseline
AI-assisted development is how we ship, not something we are piloting. Concretely, what we expect from day one:
Daily use of AI coding agents (Claude Code, Cursor, Codex or similar): required, not I tried it once"
50%+ of the code you ship is drafted by agents and reviewed by you line by line
Able to run 24 agents in parallel on isolated branches or worktrees and merge the results yourself
You maintain agent context for your projects: CLAUDE.md / AGENTS.md, custom commands, rules files. At least 1 such artifact you wrote yourself and keep current
You automate at least 1 recurring routine per monthwith an agent: reports, migrations, release checks, test generation
You can tell when the agent is wrong and throw the output away. Whatever ships, you are the engineer on the hook for it we ask about this in the interview
Nice to have
Hands-on experience with the Meta Marketing API or the LinkedIn Marketing API: breakdowns, attribution windows, async reports
Experience with a workflow orchestrator (Airflow, Prefect, Temporal). We run on Cloud Scheduler and Cloud Tasks today, and we want a view on whether that should change
The programmatic and ad-server ecosystem: CM360, DV360, StackAdapt, TikTok
Advertising metrics and how they are computed: impressions, clicks, conversions, ROAS, attribution models
Healthcare or pharma domain, especially consent and PII handling
Building agent harnesses, orchestrators or MCP servers; terminal-first workflow (tmux or equivalent)
Infrastructure as code. Our infrastructure is in Pulumi, with Terraform for the monitoring stack

Responsibilities
Own the architecture and evolution of all platform integrations: Meta Marketing API, LinkedIn Marketing API, CM360 and DV360, with StackAdapt and TikTok in flight
Lead a team of 24 engineers: code review, sprint planning, mentoring, input into hiring
Stay hands-on: you write code yourself as well as distributing it
Design and maintain the pipelines end to end: ad-platform APIs into BigQuery through Cloud Run Jobs, onward into ClickHouse through Cloud Functions and Cloud Tasks, with a large share of the transformation in BigQuery scheduled queries
Keep the smaller integrations healthy too, such as the five-minute PostgreSQL to Airtable sync
Write the FRD before an integration starts and keep the technical documentation current after it ships
Keep sync reliable: monitoring, alerting, error handling, retry and backfill strategies
Build out CI/CD, secret handling and test coverage for these services
Drive new integration projects from API evaluation to production rollout
Prioritise the integration roadmap together with the product team
Report to the CTO and agree architecture and technical standards with them

Tech stack
Python 3.11 · GCP (Cloud Run services and jobs, Cloud Functions, Cloud Scheduler, Cloud Tasks, Cloud SQL, GCS, Secret Manager) · BigQuery, ClickHouse, PostgreSQL · Docker, Cloud Build, Pulumi, Terraform · Grafana, Slack · Claude Code

What we offer
Fully remote with flexible hours, overlapping EU timezones for syncs
A direct line to the CTO: you shape the architecture and technical standards of the integration layer, and technical direction is agreed with the CTO rather than handed to you as tickets
Paid AI tooling, including Claude Code
Ad-platform APIs that fail in ways no documentation describes, at real data volumes
A small team with no bureaucracy, where decisions happen the same day