Case Studies

Research-backed RL engagements, written up after delivery. Each one is a fixed-scope production implementation — what we audited, what we built, and what shipped.

Published Engagements
Performance Engineering Tech / Inference
Inference Engine Rewrite: 100×+ Speedup in Belief Propagation
A 131× speedup. A 10× memory reduction. A full 18-property verification suite. Reimplementing a working prototype in pure C unlocked production workloads that were previously infeasible.
Stefano Ghirlanda · Aug 2, 2026 · 6 min Read →

Have an RL deployment challenge?

We work with companies from research audit through production deployment. Engagements start at $30K.

Work With Us