Finding transfer pathways in connected data.

The context

Credit Mobility Triangulator was developed by OneOrigin and sponsored by Arizona State University. It helped university administrators discover potential course equivalencies for transfer-credit evaluation. I led development of the platform and a distributed engineering team across the US and India.

The challenge

Researching transfer-credit equivalencies could take administrators days, weeks, or months. New cross-state equivalencies were often investigated only when a student requested a transcript evaluation. The challenge was to discover useful candidates proactively across a large, connected dataset and give administrators a basis for reviewing them.

The approach

The central technique was pattern-based triangulation. In a simplified example, existing A ↔ B and B ↔ C relationships could suggest an A ↔ C equivalency for administrator review. The actual patterns involved additional conditions, with confidence and scoring layered onto the recommendations.

I wrote efficient openCypher queries to match graph patterns. A clustered AWS Neptune database supported intensive graph traversal, with nightly processing across hundreds of thousands of course nodes and equivalency edges to discover matching patterns.

The work also incorporated statewide curriculum data, including transferable-credit relationships among public institutions with similar degree programs. Processing those relationships helped expand the equivalency dataset across state boundaries and surface candidates before an individual student requested an evaluation.

As technical lead, I also architected the cloud-native platform, directed the distributed team, and delivered a working demonstration to stakeholders for feedback.

The outcome

Nightly pattern discovery expanded the pool of potential equivalencies available for review. With confidence and scoring built on top of the graph analysis, the platform had the potential to bring candidate evaluation from lengthy manual research to seconds-level results, while university administrators reviewed the recommendations.

My contributions

• Led Triangulator development and provided technical direction, mentorship, and code reviews for a distributed team across the US and India.
• Wrote efficient openCypher queries for pattern-based discovery of potential course equivalencies in AWS Neptune.
• Worked on nightly graph processing across hundreds of thousands of course nodes and equivalency relationships.
• Incorporated statewide curriculum relationships to expand equivalency discovery across state boundaries.
• Architected cloud-native services using Python, Flask, SQLAlchemy, AWS Lambda, Step Functions, Neptune, PostgreSQL, and Vue.js.
• Delivered a working demonstration and gathered stakeholder feedback.