What are the responsibilities of a Senior Data Architect?
Most architect roles are either scoping or delivery. This one is both. You scope the work and see it through. Roughly half your time goes to pre-sales by sitting with retail clients during discovery, shaping what their cloud data platform should be, and defending the design in front of their C-level. The other half goes to delivery – staying close to the engineers building it, reviewing their work, and making sure the platform that ships is the one you designed.
The platforms themselves are modern omnichannel retail data stacks: cloud-native, AI-ready, and increasingly expected to power not just dashboards but machine learning (ML) models, GenAI apps, and AI agents. You will architect across the full stack, but the part of the job that's distinctive isn't the stack; it's the closeness to both the client conversation and the build.
Day to day, you will:
- Lead architecture during discovery – translating each client's business model, KPIs, and AI goals into a target-state design and implementation plan.
- Stay on as the design owner during build, reviewing pipelines, models, and BI work from data and analytics engineers, and unblocking the team when they hit hard calls.
- Architect AI-ready data foundations – the semantic layers, feature stores, vector databases, and embeddings pipelines that ML models, GenAI apps, and AI agents need to work reliably.
- Design end-to-end cloud platforms for retail clients: ingestion, storage, transformation, modeling (dimensional, Data Vault, One Big Table, or whatever fits), BI, and CDP.
- Build engineering rigor in by default – automated pipelines, infrastructure-as-code, CI/CD, testing, and observability – so each platform gets cheaper and more reliable over time, not the opposite.
- Co-sell with Business Developers, Key Account Managers, and Strategists – shaping proposals, defending estimates, and presenting architectures to client stakeholders who range from skeptical CTOs to non-technical executives.
Platforms and tools you'll work with:
- Cloud: AWS, GCP, Microsoft Fabric
- BI: Looker Studio, Power BI, Tableau
- CDP: Segment, mParticle, Bloomreach, RudderStack, Snowflake-native
- AI coding assistants: Claude Code, Cursor, Copilot, in-warehouse LLM features
- Transformation & modeling: SQL, dbt, Python
What we expect from you
We hire for mindset and impact first. You should hold a clear technical opinion, communicate well with both engineers and executives, and stay practical about what each client needs.
Mindset
- Automation-first. You reach for infrastructure-as-code, CI/CD, testing, and observability – not runbooks and Slack pings. A good platform needs less babysitting over time, not more.
- You know the trade-offs, not just the brand names. You can say why BigQuery fits one client, and Snowflake fits the next.
- You actually use AI coding assistants in your work – generating SQL, dbt, and Python, reviewing designs, drafting docs – not just opening Cursor once and closing it.
- You can defend an architecture to a skeptical CTO and a non-technical executive in the same room. You run workshops and challenge assumptions without putting people on the defensive.
Experience
- 5+ years in data architecture, engineering leadership, or analytics engineering, with at least three cloud platforms designed and shipped – ideally in retail or e-commerce.
- Production experience across the stack below. You don't need every tool, but you should be deep in at least one per category.
- A practical sense of what ML, GenAI, and AI agents need from a data platform – feature stores, embeddings, vector databases, RAG, semantic layers, governance, lineage. We want someone who can learn fast across these, not someone who's shipped all seven.
- Working knowledge of the BI and activation side – dimensional modeling, semantic layers, CDPs, identity resolution, consent management, and downstream activation into ad, CRM, and personalization tools.
scandiweb is an international company; therefore, confident spoken and written English is essential.
