InTowards AIbyEbrahimi·Jul 9Physics-Informed AI, Part III: From Post-Hoc Checker to Differentiable Physics HeadPart II checked physics after generation. Part III puts the residual inside the training loop with a language-conditioned numerical head.A response icon2A response icon2
InTowards AIbyEbrahimi·Jun 29Physics-Informed AI: Fine-Tuning an LLM to Speak Engineering While the Checker Owns the PhysicsPart 2 of a practical series — The model learns the schema. The checker owns the physics. Those are different jobs.
InTowards AIbyEbrahimi·May 2Physics-Informed AI: Why LLMs Need Solvers, Constraints, and Physical LawsPart 1 of practical series — An engineering view of how LLMs, physics-informed ML, and numerical solvers can work together without…
InTowards AIbyEbrahimi·Apr 5PINNs and Neural Operators: Two Competing Visions of Scientific AIOne embeds the governing equation into training. The other learns a reusable map from problem setup to solution.A response icon1A response icon1
InTowards AIbyEbrahimi·Apr 4Stop Reinventing the Solver StackHow engineering teams in simulation and ML waste years rebuilding what already exists, and what to do instead