Design an SAT-Based Verifier for an Autonomous-Vehicle Test Lab
Overview
What this challenge is about.
Model a simplified four-way intersection: agent positions, lights, and discrete time steps. Define 5 safety properties in propositional logic (e.g., 'no two agents in the intersection at conflicting times'). Build a Python harness that takes a candidate policy and a bounded horizon, encodes the joint trajectories + property negations as a CNF (conjunctive normal form) formula, and dispatches to an open-source SAT solver (MiniSat or Glucose via PySAT). If the solver finds a model, you have a counterexample; if not, the policy is safe within the horizon. Document the encoding, run the verifier on 3 candidate policies, and write a 4-page report on scope, limits, and a 6-month investment recommendation.
The Brief
What you'll do, and what you'll demonstrate.
Build a bounded SAT-based safety verifier for candidate AV policies on a stylized intersection model.
Earning criteria — what you'll demonstrate
- Encode a discrete-time safety problem as propositional logic
- Use a SAT solver as a verification back-end via PySAT
- Interpret SAT solver outputs as safety counterexamples
- Communicate the scope and limits of bounded verification
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
AI Engineer
Implementing a SAT-based verifier and integrating a solver is the kind of foundational AI engineering that AV safety teams actively hire for.
This challenge sharpens
- sat-solving
- formal-verification
- python
AI Safety Researcher
Bounded safety verification of policies is a transferable AI safety craft, applicable from AVs to general agent governance.
This challenge sharpens
- safety-engineering
- formal-verification
- logical-inference
AI Solutions Architect
Defending a verification investment against a simulation-only alternative is the architect's bridge between tools and team strategy.
This challenge sharpens
- safety-engineering
- knowledge-representation
- formal-verification