Skip to content

Data Engineer - Contractor Role

Job Introduction

About Tribepad

At Tribepad, we put People first, Product second and Profit third. Our recruitment technology is used by employers such as Greggs, Northumberland County Council, the Met Office and DFS.

We're on a mission to make recruitment faster, fairer and better for everyone. Along the way, we've earned these accreditations:

  • Certified B Corporation

  • ISO 9001, ISO 14001 and ISO 27001 certified

The Role

Six-month Contractor role, full-time from January 2027, reporting to the CPTO.

Mostly remote - Sheffield two or three days a month for planning and team days. Standard day rate, paid monthly.

We are building Tribepad's new recruitment platform. Reporting and AI are built in from the start rather than bolted on. Right now, there is no data platform behind it. You will stand one up from scratch, get the first customer-facing analytics and AI answers running on it, and leave behind a team that can run and extend it without you. Half the job is the platform. The other half is upskilling our existing engineers in modern data pipelines, modelling, testing and tooling, so the capability stays with Tribepad when the contract ends.

This isn’t a DBA or SQL analyst role. It’s for a hands-on lead who has built a production warehouse before and can teach others to do it.

What you'll do

By the end of six months, we expect:

  • A production data warehouse built on a greenfield basis: environments, schema layout, role hierarchy and access model, naming conventions, CI/CD for data changes, all as code and reproducible.

  • Change data capture from the PostgreSQL database into the warehouse, handling schema changes, deletes, late-arriving data and JSONB.

  • A dimensional and semantic model where a metric is defined once, so time to hire means the same thing in every dashboard, report and AI answer.

  • Multi-tenant isolation in the warehouse that mirrors  Postgres row-level security. A tenant can never see another tenant's data through a dashboard, an AI-generated query or a data share.

  • The first customer-facing reporting is running on the warehouse, and a semantic layer that an LLM can answer questions over, tested against a golden set of real recruiter questions.

  • Data quality and observability in place: freshness, volume and schema tests, alerting and lineage.

  • Privacy engineering done properly for candidate data: UK and EU GDPR, right to erasure propagating into the warehouse, retention, pseudonymisation of special category data such as diversity monitoring.

  • A data strategy written down and agreed, covering what we build, buy and defer beyond the six months.

  • The existing team upskilled in modern data engineering: pipelines, dimensional modelling, dbt-style transformation, testing, observability and infrastructure as code. Pairing, code review and short working sessions, not classroom training. 

The last point matters as much as the first. We are hiring you to make the team better, not to be the only person who understands the data platform.

Before you apply:

  • You are available full-time from 1 January 2027 for six months, with no overlapping contracts.

  • You are operating through a limited company and set up to work outside IR35: own kit, own methods, substitution clause acceptable.

  • You are UK-based with the right to work, able to get to Sheffield two or three days a month and work UK hours.

  • You've stood up a production data warehouse from scratch at least once and can walk us through the decisions made and what you'd do differently. Extending someone else's warehouse does not count.

  • You can provide references from a product or engineering team they were embedded in, not just an agency client.

Experience

Our current thinking is Snowflake. If you have a strong case for something else, make it, but you need the depth below in whatever you propose.

  • Setting up a data warehouse from scratch. Account and environment structure (dev, test, prod), database and schema layout, role hierarchy and access model, warehouse sizing and resource monitors, naming conventions, CI/CD for data changes, and the ingestion and transformation foundations everything else is built on.

  • Dimensional and semantic modelling. Kimball-style facts and dimensions, slowly changing dimensions, and defining metrics once so they mean the same thing everywhere.

  • Snowflake in depth. Warehouse sizing, clustering, Dynamic Tables or Streams and Tasks, RBAC design, row access policies, masking policies, secure data sharing and cost monitoring.

  • Infrastructure as code for the warehouse. Accounts, roles, warehouses and policies managed through Terraform or equivalent, so the platform is reproducible and changes are reviewed.

  • Semantic layer for AI. Snowflake Semantic Views and Cortex Analyst or equivalent: semantic models with descriptions, synonyms, verified queries and custom instructions, and evaluating accuracy against a test set of real questions.

  • dbt (or equivalent) as the transformation layer. Tested, version-controlled, documented models with CI. Publishing dbt metrics into semantic objects is a plus.

  • Change data capture from PostgreSQL. Logical replication, schema changes, deletes, late-arriving data and JSONB columns.

  • Multi-tenant data isolation. Mirroring row-level security from the application database so no tenant can see another's data by any route.

  • Strong SQL and Python. Python for orchestration, data quality checks and glue code.

  • Data quality and observability. Freshness, volume and schema tests, alerting and lineage, so problems are caught before customers see them.

  • Privacy engineering for personal data. UK and EU GDPR, right to erasure propagating into the warehouse, retention policies, and pseudonymisation of special category data.

  • Has coached or mentored engineers and can point to someone who got better because of it.

  • Uses AI tools in their own engineering work already and can show how.

You will sit in one of three cross-functional teams, working in two-week sprints with a sprint demo every fortnight. The team is small and senior enough that you will be making the architectural calls, not waiting for them. AI is native in our product, so every data problem gets the question "what does an agent do with this" before we design a report.

Apply

This website is using cookies to improve your browsing experience. Tracking cookies are enabled but these do not collect personal or sensitive data. If you prefer for this not to be collected, please choose to turn cookies off below. Read more about cookies.