AI Data Engineer

SUMMARY

Build the data foundation that grounds AI agents on facts they can trust. You turn raw, messy, multi-source data — structured feeds and unstructured content alike — into canonical, provenance-tagged, SLA-fresh Gold data that the Knowledge & Intelligence layer serves. Working with business leaders and internal customers, you set the design standards and assurance processes that keep the pipeline compatible, operable, and traceable end to end. This is hands-on data engineering where correctness and lineage aren't nice-to-haves — they're the product. You develop fundamentally new approaches to pull meaning out of data, and every record you land answers the question “where did this come from, and can we trust it?”

What You'll Own

• The medallion pipeline. Take landed data through Bronze → Silver → Gold — conforming structured sources and extracted content to Gold within its freshness SLA, with traceable lineage at every hop.
• Entity resolution. Own the resolve gate that turns many sources into one canonical identity — the join everything downstream depends on.
• Model-based extraction lanes. Build the NLP- and LLM-assisted lanes that pull structure out of unstructured content across many content types — HTML, PDF, DAM assets, and more — tagged as evidence, not fact.
• Provenance and trust. Tag provenance and trust tier on every record. Quarantine what fails the gate with its reason attached, rather than silently dropping it.
• Performance at scale. Keep high-volume data in the lakehouse — and the served layer — lean and fast.

Why This Role?

• You own the layer everything else is built on. If the grounded facts are right, every agent that reads them is more trustworthy; if they're wrong, nothing downstream can save it — so this is where the leverage really is.
• Genuinely hard, genuinely new work: production-grade extraction from unstructured content, resolution at scale, and lineage you can defend — not another ETL backlog.
• A pod where provenance and trust are first-class engineering concerns, with the mandate to build the gates that enforce them.

Essential Requirements:

• Strong model expertise for unstructured data — NLP and LLM-based extraction, document understanding, embeddings. This is a core need for the pod.
• Strong PySpark and lakehouse / medallion experience in production.
• Advanced SQL — complex transformations, window functions, and query tuning.
• Proven experience handling data at high volume — distributed processing over large datasets, with real attention to cost, partitioning, and performance.
• Lakehouse table formats — Delta Lake, Iceberg, or Hudi.
• A cloud data platform — Databricks or Snowflake.
• Entity resolution / record linkage at scale.
• Airflow or comparable orchestration.
• A provenance mindset — “where did this come from” is a first-class column to you.

Benefits and Perks of working at Dell Technologies

Your life. Your health. Supported by your benefits. You can explore the overall benefits experience that awaits you as a Dell Technologies team member — right now at MyWellatDell.com