Safety Kernel Insights

Practical architectures, proofs, and design decisions from building deterministic safety kernels.

This section offers technical perspectives drawn from building deterministic safety kernels and related systems. Articles focus on architectures, formal methods, safety certification, and the engineering principles that matter when reliability, reproducibility, and verifiable behaviour are essential.

Systems Architecture January 26, 2026 20:05

The Init-Update-Status-Reset Pattern: O(1) Guarantees for Safety Monitors

A four-function interface that enables static analysis, bounded resources, and compositional verification

10 min read →
Formal Methods January 26, 2026 19:00

Contracts as Documentation: Why Comments Lie and Code Doesn't

How preconditions, postconditions, and invariants become living documentation

8 min read →
Safety-Critical Programming January 26, 2026 18:30

Why 'Hello World' Fails Safety-Critical Engineers

Traditional C tutorials teach habits that certification auditors reject

8 min read →
Announcements January 24, 2026 22:31

C From Scratch: A New Approach to Learning Safety-Critical C

Why proving code correct before writing it changes everything

7 min read →
Deterministic Computing January 20, 2026 22:00

Stochastic Rounding Without the Stochastic

How PRNG-controlled rounding can provide regularisation benefits deterministically

6 min read →
Deterministic Computing January 20, 2026 20:45

Cross-Platform Bit-Identity: From Theory to 7 Matching Hashes

The practical journey of verifying deterministic ML across platforms

7 min read →
Deterministic Computing January 20, 2026 20:24

The Feistel Shuffle: Deterministic Data Ordering Without Randomness

How cycle-walking Feistel networks can provide reproducible shuffling for ML training

6 min read →
Deterministic Computing January 20, 2026 19:30

Merkle Chains for ML Audit Trails

How cryptographic hash chains can make every training step verifiable

7 min read →
Deterministic Computing January 20, 2026 19:00

Round-to-Nearest-Even: The Rounding Mode That Makes Determinism Possible

Why banker's rounding matters for bit-identical machine learning

6 min read →

10 of 34 articles