Graduate mentor's supervisor: Grant Weddell
Research Problem
Finding and combining hidden data in modern data lakes currently relies on black-box AI tools or trial-and-error, offering no guarantees of correctness. This project introduces SQLP-DL, a brand-new, formal query language. Solving this transforms data discovery from an unpredictable guessing game into a reliable science, allowing programmers to safely combine, optimize, and reuse data pipelines without errors.
Importance:
You get to help invent a new database language! Just like standard SQL revolutionized databases decades ago, SQLP-DL introduces new commands like UNIONABLE and JOINABLE as fundamental math rules. You will be building the very foundations that future data platforms will run on.
Your team of 3–4 will collaborate on building this language framework:
- Short-term: Define the syntax rules for the new commands and write a basic code parser.
- Medium-term: Build a lightweight execution engine to run simple SQLP-DL queries over a mock data lake.
- Longer-term: Explore basic query optimization strategies to make the language run faster.