Middleware Engineer

Position Overview

We are hiring a Middleware Engineer to own the layer between the ad platforms and our AI: the integrations, the pipelines, the queues and the data. Our AI agents are only as good as the data under them. That data comes from Meta and Google Ads, every day, for every brand, through APIs that rate-limit, tokens that expire, webhooks that fire twice and jobs that time out. If ingestion is wrong, every insight, report and agent answer on top of it is wrong. Your job is to make it boring, correct and fast.

The Hard Problems

• Pull ads, creatives and metrics daily for every brand and every ad account, through APIs we do not control
• Make every job safe to run twice: no duplicates, no silent gaps
• Keep each brand's data, files and tokens fully separate, on every route and every job
• Keep one definition of ROAS, CPC, CTR and the rest, used by every screen and every agent

Core Responsibilities

• Platform integrations: own the Meta and Google Ads API clients (token lifecycle, rate limits, pagination, retries), built so a new platform plugs in without forking the code
• Ingestion, webhooks and jobs: own the daily ingestion pipeline, webhook intake with verification and dedup, and background jobs on Celery, Redis and SQS, with queues split so a chat request never waits behind bulk ingestion
• Data layer and metric correctness: own one MongoDB database per brand (indexes, upserts, uniqueness, migrations), Redis caching that clears when data changes, and one shared definition of every metric
• Auth and tenancy: own JWT middleware on every route, tenant checks on every read, write and file key, and secrets in a secrets store, not in code
• Ship, run and observe: own Docker, AWS ECS and GitHub Actions with tests that gate the deploy and one-step rollback, plus logs, traces and alerts so we know a brand's data is stale before they do; own high-level and low-level design for this layer

Required Qualifications

• 3-5 years of total experience in backend engineering; Python and FastAPI (or a similar async framework) in production
• Built and ran third-party API integrations at scale: rate limits, OAuth tokens, pagination, partial failures
• Built data pipelines or background job systems in production: Celery, SQS, Kafka or similar
• Idempotency and retries that can be explained with a real incident (the duplicate you found)
• MongoDB or another database in depth: indexes, aggregation pipelines, query plans, schema changes on live data
• Multi-tenant systems: how you kept one customer's data away from another's
• Caching with Redis or similar: when to cache, when to clear, and what went wrong
• Cloud and CI/CD: AWS (ECS, S3, SQS), Docker, GitHub Actions or equivalent
• Security basics: authn vs authz, secret handling, SSRF, injection; fixed in code, not in a checklist
• Written design docs; led at least a small team or a large project end to end

Good to Have

• Meta Marketing API or Google Ads API experience
• Webhook systems at volume (signature checks, replay, ordering)
• Observability (OpenTelemetry, New Relic, Sentry)
• Work next to an AI or LLM team (clean data for agents, MCP tools)
• Ad tech, martech or analytics products

Why Join

Every insight and every agent answer rests on the data layer you build. This is an early role in a company built from the ground up, and what you build becomes the foundation others build on.