Clear, practical writing at the intersection of AI, data science and quantum-safe security — the same topics we work on every day, explained for the people who have to make decisions about them.
Our notes span two horizons. On one side, applied AI and data engineering: how to move models into production, mine unstructured data, and build systems teams can trust. On the other, the shift to quantum-safe cryptography: why organisations should start with cryptographic discovery and a CBOM, and how crypto-agility makes migration manageable rather than disruptive. New quantum-safe pieces are in preparation and marked accordingly.
You cannot migrate what you cannot see. We look at why cryptographic discovery and a CBOM have to come first, and how designing for crypto-agility turns a one-off scramble into a repeatable, low-risk process.
A CBOM is to cryptography what an SBOM is to software. We explain what it captures — algorithms, key lengths, certificates, protocols and their dependencies — and why that structured inventory is the foundation for any controlled migration.
A practical look at the newly standardised algorithms — ML-KEM and ML-DSA — where hybrid classical-plus-PQC schemes make sense, and what "harvest now, decrypt later" really means for data that must stay secret for a decade or more.
Finding cryptography across a large codebase is a classification problem at heart. We explore where machine learning can accelerate discovery and dependency analysis — and where human review remains essential. Research direction.
How neural networks and evolutionary algorithms are accelerating the discovery of high-strength, lightweight structures — and what that means for teams bridging computational modelling and physical R&D.
Why keyword queries fail on real documents, and how retrieval-augmented generation (RAG) is changing the way teams search technical literature, contracts and internal knowledge — safely and with citations.
A practical guide to wiring together databases, message brokers and warehouses so that machine-learning predictions can stream continuously — without downtime, drift or data loss.
How production lines use anomaly-detection models to spot deviations early and prevent unplanned outages — and why standardising data before adding AI is what actually makes it stick.
What quantum computing may and may not change for machine learning and cryptography — a grounded look at timelines, real risks like "harvest now, decrypt later", and the sensible steps to take today.
Quantum-safe articles marked "Coming soon" are in preparation; the remaining posts are retained and updated from our previous site.