Data Engineer

About Lumo

Lumo is an early-stage SaaS platform for the TikTok Shop affiliate ecosystem. We help brands discover, score and activate the right creators at scale, in the fast-growing model of community commerce: affiliate-led selling built on high volumes of smaller, authentic creators rather than paid ads or a handful of large influencers. We are building the data and intelligence foundation that powers creator matching and content guidance. This role owns that foundation.

The Role

You will build and scale the creator data platform that everything else runs on. You will own the pipelines that acquire, clean, structure and serve data across the product, and you will build the single source of truth our analytics and machine-learning models depend on. You will work directly with the founder and core engineering team, with real autonomy over how the data layer is built.
This is a foundational hire. The matching model and the content-intelligence model cannot be trained until this layer is trustworthy, current and well structured. Reliability is the job, not a nice-to-have.

*What You'll Do
*

• Build and maintain ingestion pipelines that continuously collect creator, content and performance data.
• Develop and manage API integrations with platforms and third-party providers (TikTok and TikTok Shop, Instagram, YouTube, Shopify, and creator-data vendors).
• Implement web scraping where appropriate, with rate-limit and anti-bot resilience, retries with backoff, and alerting when a source silently breaks.
• Design a unified data model linking creators, brands, campaigns, content, products and performance, as a genuine single source of truth.
• Solve cross-platform entity resolution: match the same creator across TikTok, Instagram and YouTube where there is no shared key.
• Build append-only, point-in-time snapshotting of creator and campaign performance, so historical depth accrues despite platform history limits.
• Build cleaning, validation, standardisation and enrichment processes for messy, real-world data.
• Build the ground-truth labelling infrastructure that links content and campaign signals to measured outcomes.
• Stand up and maintain the warehouse or feature store and a scalable storage layer, including large media where video ingestion is in scope.
• Implement event tracking that captures in-product behaviour as structured, proprietary first-party data.
• Establish data-quality monitoring, freshness SLAs and alerting so the platform runs on trusted data.
• Document pipelines, schema and runbooks as you go.

*Must-Have Skills
*

Must have
• Three or more years in data engineering or a closely related role.
• Strong Python for ETL and ELT.
• Strong SQL and relational design (PostgreSQL preferred).
• Proven data modelling and source-of-truth schema design.
• REST APIs and third-party integrations.
• Pipeline orchestration (Airflow, Prefect or similar), with idempotent, incremental design, deduplication, backfill and schema evolution.
• Cloud experience (AWS preferred).
• Comfort with web scraping and inconsistent real-world data.

*Nice to have
*

• Data warehouse experience (BigQuery, Snowflake or similar).
• Feature store experience, given the ML work that follows.
• Social-media, creator or marketing data.
• Data-quality tooling and monitoring.
• Vector storage and embeddings (pgvector), helpful as the models land.
• Media pipelines: frames, audio and transcripts at scale.**

What We're Looking For

Someone who takes ownership, works well with limited hand-holding, and cares about building things that last rather than quick hacks. You'll be laying the foundation that the whole platform's intelligence depends on, so reliability and clean structure matter to you. Comfortable in an early-stage, fast-moving environment where priorities shift.