Use Cases
Fourteen real workloads, each pushing linaldb into a genuinely different domain: no
synthetic toy examples, no cherry-picked results. Every one is adapted from a real,
fully-executed notebook in the linal-hub
suite, cross-checked against numpy (or, where applicable, known analytical facts)
at every step. Seven real engine bugs were found (and fixed) building these.
| Use case | Domain |
|---|---|
| Product Catalog Search | Hybrid tables + similarity search |
| Leukemia Subtype Classification | Genomics |
| Gravitational-Wave Detection | Astrophysics |
| Convolution via MATMUL | Computer vision |
| Lineage & Linear Algebra | Provenance |
| LLM Weight Compression | ML / NLP |
| Single-Cell PCA | Genomics |
| Production Network Risk | Economics |
| Movie Recommender | Recommender systems |
| Manufacturing Quality Control | Industrial / process engineering |
| Dynamical Systems & Chaos | Pure mathematics |
| Climate Science Reanalysis | Climate science |
| Astronomy Catalog Hybrid Search | Astronomy |
| Control Systems Stability | Control engineering |
Each page links to the full source notebook and a “Try it in the Playground” deep link that pre-loads a small, self-contained version of that use case’s real query.

