Backend Engineer

O projekcie:
About Little EngineLittle Engine is building the AI-native back office for solo service providers — personal trainers, stylists, tutors, massage therapists, dog walkers, and more. Our mobile app automates scheduling, payments, and client communication — entirely over SMS — replacing a mess of Venmo, Notepad, and texting with a clean, intelligent assistant that works behind the scenes.

We closed a multi-million dollar pre-seed.

The team includes repeat founders who’ve built and sold award-winning apps — including one of Apple’s Best Apps of the Year.

Wymagania:
What you'll work on
React Native app for solo entrepreneurs in the US — a back-office tool that helps them manage scheduling, payments, and client communication from their phone. The app is their daily driver for running the business, so reliability, performance, and polish matter.

The stack

- Language & framework: Python 3.12, FastAPI, Pydantic v2, uv (deps/lockfile)- Data & persistence: PostgreSQL, SQLAlchemy 2 (async) with asyncpg + psycopg 3, Alembic migrations, Redis- Async & jobs: Celery (workers + beat) on Redis, uvloop, tenacity for retries/backoff- AI / LLM: Anthropic Claude, Model Context Protocol (MCP) tool server via FastMCP, relevance-classifier agents- Backend integrations: Twilio (SMS, Voice, Verify, messaging), Stripe, Firebase (Firestore, Cloud Messaging), Google Calendar/OAuth, Cloudflare R2 (S3-compatible via boto3) + CDN, Amplitude, Sentry- Tooling: Docker (multi-stage) + Compose, GitHub Actions CI/CD, GCP (Cloud Run, Artifact Registry, Workload Identity Federation, Cloud SQL, Memorystore), Ruff, mypy + ty, pre-commit, pytest (asyncio/xdist/cov) + Dockerized E2E harness

What we're looking for
- Senior-level experience shipping and operating production Python backend services (typically 5+ years, but we care more about depth than tenure)- AI coding agents are a normal part of the daily workflow — we delegate, review, and steer agent output rather than treating it as a novelty- Strong Python fundamentals — async/await, typing — and sound REST API design instincts- You've owned third-party integrations end-to-end, especially payments (Stripe) and calendars (Google Calendar/OAuth) — and you know these domains are where the hard problems hide: webhooks, idempotency, retries, reconciliation, time zones, and recurring events- Hands-on with FastAPI, SQLAlchemy (async), and Pydantic; comfortable working inside a Clean / Hexagonal architecture and respecting its boundaries- Fluent with background processing on Celery + Redis — webhook-driven flows, idempotent task design, and getting eventual-consistency right- Strong with PostgreSQL — schema design, migrations (Alembic), and reasoning about query behavior when the work calls for it- Experience wiring up other platform SDKs — Twilio, Firebase (Firestore, Cloud Messaging), Cloudflare R2/S3 (or equivalents)- Comfortable with Docker, GitHub Actions CI/CD, and deploying to the cloud (GCP / Cloud Run or equivalent)- A real testing discipline — unit, integration, and end-to-end — because money and bookings have to be correct- Comfortable working across a time-zone gap with a US-based product team

How we work- The engineering team is based in Poland; the product team is in the US. Expect meaningful overlap with US hours for product conversations, and asynchronous-by-default work the rest of the time.- Small team, startup pace — scope shifts, priorities move, and shipping matters more than process- You'll work directly with product and backend, not through a chain of intermediaries — expect to be in scoping conversations, push back on requirements, and shape APIs alongside the people building them- Engineers own areas of the product end-to-end — you're accountable for a slice of functionality from technical design through production behavior, not just the tickets assigned to you

Application note
We use AI tools ourselves, so AI-assisted applications are fine. What we do not want is a generic AI-written summary of this job post.

Please include a short section titled “Boring production detail” and describe one specific issue you have personally dealt with in production: a webhook retry, payment reconciliation problem, OAuth/token edge case, calendar time-zone bug, recurring-event sync issue, slow query, unsafe migration, or background-job duplication.

Keep it concrete: what broke, how you detected it, what you changed, and what you would monitor afterward.

Codzienne zadania:
- Build, ship, and maintain features in the backend platform — from technical design through deployment and production monitoring (Sentry)
- Own a slice of the product end-to-end: scope the work with product, design the implementation, ship it, and stay accountable for how it behaves in production
- Work directly with the US-based product team in scoping conversations — push back on requirements, surface trade-offs early, and shape what gets built rather than just implementing tickets
- Own the API contract — collaborate with mobile and web engineers on endpoint design, and shape clean, well-versioned interfaces that are easy for clients to consume
- Lead our payments and calendar integrations — Stripe (charges, refunds, disputes, webhooks) and Google Calendar (OAuth, sync, recurring events, time zones) — and own the correctness of money and bookings
- Design resilient async workflows — Celery jobs and webhook handlers that are idempotent, retry-safe, and correct under eventual consistency
- Hold the line on performance and reliability — profile slow queries, add the right indexes, keep async paths non-blocking, and make sure background jobs and webhooks scale
- Drop down to raw SQL or tune SQLAlchemy when the ORM isn't enough — reason about transactions, locking, and running migrations safely against production data
- Integrate and maintain third-party SDKs (Stripe, Google Calendar, Twilio, Firebase, Cloudflare R2) — including the unglamorous parts: version upgrades, auth/token flows, edge cases, reconciliation, and production incidents
- Occasionally reach into the client apps (React Native / React.js) when an integration needs glue on both ends
- Keep the codebase healthy — type checking (mypy/ty), Clean Architecture boundaries, reusable use cases and ports, and sensible domain modeling