Data Intern

We're looking for a Geospatial Data Engineering Intern to help build and scale our geospatial data infrastructure over the summer. This role is designed for a junior or senior undergraduate with at least one prior internship under their belt, strong data engineering instincts, and the team awareness to ship work in a shared production codebase. You'll get hands‑on experience building pipelines that ingest, transform, and serve geospatial data with exposure to AI agent tooling along the way.

This role begins as a full‑time, 3-month summer internship and then continues part‑time through September
.

You'll work directly with our data team, contributing to operational infrastructure that powers geospatial analysis and decision‑making across the organization. Your primary focus will be building reliable, well‑documented data pipelines with a geospatial backbone, while getting meaningful exposure to applied AI systems and helping us complete an in‑flight migration from Airflow 2 to Airflow 3.

You’ll spend the majority of your time on geospatial data engineering, with supporting work in geospatial analysis and applied AI.

• Build and improve Airflow ELT pipelines that ingest, transform, and serve geospatial datasets at scale, working across both our Airflow 2 and Airflow 3 repositories and actively assisting with the Airflow 2 to 3 migration, including porting DAGs, validating parity, and helping retire legacy pipelines
• Write clean, type‑hinted Python and well‑structured SQL, including geospatial operations (PostGIS, spatial joins, CRS management) against Athena (Trino), Postgre
  
  SQL, Redshift and DuckDB
• Develop modular dbt models with semantic layer definitions and documented business logic for geospatial tables
• Contribute to data quality systems, including schema validation, freshness monitoring, and spatial integrity checks
• Support Data Hub adoption through schema documentation, lineage tracking, and metadata management for geospatial assets
• Triage failing DAG runs, read Airflow task logs, and own fixes end‑to‑end
• Communicate progress through documentation, code reviews, and regular updates

• Contribute to research‑oriented analyses such as tree canopy classification, network resiliency analysis, and spatial feature extraction
• Design and document reproducible analytical workflows that feed into production pipelines
• Translate complex geospatial methods into clear, accessible outputs for non‑technical stakeholders
• Share learnings on emerging GeoAI methods and geospatial tooling with the team

• Assist with building data agents using tools like Lang Graph, Lang Chain, or Bedrock Agent Core
• Support development and iteration on pipelines and text‑to‑SQL approaches for natural‑language data access
• Contribute to MCP server development and agent evaluation as needed
• Document agent failure modes and help refine prompts based on feedback

• Currently a junior or senior undergraduate (or higher) in Computer Science, Data Science, GIS, Geospatial Engineering, Software Engineering, or a related field
• At least one prior internship (or equivalent team‑based engineering experience); you've shipped code in a shared repo, taken a code review, and worked a ticket end‑to‑end
• Available to work full‑time for 3 months during the summer, then part‑time through the fall semester
• Strong fundamentals in Python, including classes, inheritance, decorators, type hints, and explicit imports
• Strong fundamentals in SQL: joins, CTEs, window functions, and aggregations
• Comfortable working in Git/Git Hub with a dev → main PR‑to‑deploy workflow
• Comfortable on the Unix command line or eager to learn (bash, navigating a file system, running scripts)
• Familiarity with geospatial concepts (CRS, spatial joins, indexing) and tooling such as PostGIS, Geo Pandas, or QGIS is a plus
• Exposure to AWS or similar cloud providers; ideally S3, IAM, Athena, Glue, ECS, or Redshift
• Experience with Airflow or similar orchestration tools is a plus (or strong eagerness to learn quickly. You'll be ramping on two versions in parallel and…