Back-end Engineer: Connectors & Integrations

Strattum goes live across many clients at the same time — and each one is a different world. Different ERPs, different databases, different APIs, schemas that never repeat. For the platform to work, all that data has to come in reliably, standardized, and versioned.

We're looking for a Back-end Engineer to build and maintain our connectors, the integrations that link each client's base to the platform. Every new client is a new connector (or an adaptation of an existing one), and that integration layer is what holds everything else up. It's real back-end work: dealing with unpredictable sources, taming the chaos, ensuring idempotency and resilience, and keeping it all running as the client base grows.

What you'll do:

• Build connectors end to end — from source extraction (database, API, ERP) to clean, standardized data ready for the platform to consume
• Handle the diversity of clients — PostgreSQL, SQL Server, TOTVS, REST APIs, and whatever else shows up
• Write cleaning and standardization transforms — normalize messy schemas into a consistent model (dedup, emails/phones, external keys, N:M relationships)
• Ensure reliability — idempotent, incremental (watermark) pipelines with error handling and reprocessing
• Maintain and evolve existing connectors — sources change, APIs break, schemas evolve. Maintenance is half the game
• Orchestrate execution — scheduling, monitoring, and observability of the pipelines

On language: our stack is Python (dlt, Prefect, FastAPI) — but we don't require prior Python. In the AI era, the language is syntax. What you can't outsource to the model is back-end reasoning: understanding a data source, deciding how to model it, ensuring reliability, debugging a pipeline that broke at 3 a.m. Bring back-end experience from any stack (Java, C#, Go, Node…) and you'll pick up Python fast.

Required:

• Solid back-end experience — data modeling, integrations, idempotency, concurrency, and failure handling
• Comfort with databases and SQL — you'll talk to many, in many ways
• A mindset for handling chaos — real data sources are messy; you normalize without complaining
• Autonomy — take a new client, understand the source, ship a working connector

Nice to have:

• ETL/ELT, data pipelines or data engineering (dlt, Airbyte, Fivetran, dbt, Airflow, Prefect)
• Having integrated ERPs or legacy systems (TOTVS, SAP, Protheus and the like)
• Knowledge of data modeling and data quality
• Having used Claude Code or AI codegen tools day to day

How to apply: record a video of up to 5 min showing some ETL code you've written, explaining what you did and why. Paste the video link in the application form.