Training Course Professional

PQC Protection for AI Model Weights and Training Infrastructure

Applying post-quantum cryptography to protect model weights, training data, and inference endpoints.

Full day (6 hours)

Certificate on completion

Enquire About This Course View Curriculum

Proud to recommend our expert members

Course Overview

For MLOps engineers and AI platform security teams. Covers quantum cryptographic exposure of model weights, training checkpoints, NCCL/MPI cluster communications, model registry signing, and PQC migration using FIPS 203/204/205 for distributed AI training.

Curriculum

Course content is structured into sessions below. Each session is configurable to your organisation's environment, threat model, and objectives.

#	Session	Topics
1	AI Training Infrastructure and Its Cryptographic Dependencies Where cryptography protects model weights, checkpoints, and cluster communications
2	Model Artefact Security Under Quantum Threat Weights, checkpoints, and model registry integrity	Cryptographic signing of model weights and training checkpoints: RSA/ECDSA vulnerability to Shor's algorithm Model registry integrity: container image signing (cosign/Notary v2) and SBOM attestation exposure Harvest-now-decrypt-later risk for encrypted model artefacts stored in cloud object stores
Break
3	Distributed Training Cluster Communications NCCL, MPI, and GPU-to-GPU cryptographic exposure	NCCL AllReduce and Ring-AllReduce: TLS/encryption status in multi-node GPU clusters (NVIDIA DGX, cloud GPU instances) MPI collective communications: OpenMPI and MPICH transport security under quantum threat Parameter server architectures: gRPC TLS and key exchange for gradient aggregation
4	Interactive Demonstration: Model Security Cryptographic Audit Full-day format only	Mapping cryptographic dependencies across a distributed training pipeline (data ingest, preprocessing, training cluster, checkpoint storage, model registry) Evaluating NCCL encryption configuration and identifying unencrypted gradient flows Demonstrating ML-DSA model signing and verification workflow for checkpoint integrity
Break
5	PQC Migration for AI Training Infrastructure FIPS 203/204/205 applied to ML workflows	ML-DSA-65/87 for model weight signing: signature size impact on model registry throughput ML-KEM-768 for encrypted checkpoint storage: key encapsulation for cloud object store encryption SLH-DSA for long-lived model provenance: stateless hash-based signatures for multi-year model lineage
6	Compliance and Vendor Landscape NIST AI RMF, EU AI Act, and platform readiness	NIST AI RMF Map 1.5/Measure 2.6: cryptographic security requirements for AI model integrity EU AI Act Article 15: accuracy, robustness, and cybersecurity requirements for high-risk AI systems Cloud GPU platform PQC readiness: NVIDIA, AWS, Azure, GCP encryption and signing capabilities
7	Q&A and Migration Planning

Designed and Delivered By

Training courses are designed and delivered by QSECDEF in collaboration with subject specialists. All instructors have direct experience in quantum technologies and security.

QSECDEF brings expertise in post-quantum cryptography, quantum computing strategy, and defence-grade security assessment. Our membership spans 900+ organisations and 1,500+ professionals working at the intersection of quantum technologies and critical infrastructure security.

Enquire About This Course

Get in touch to discuss scheduling, group pricing, and how this course can be configured to your organisation's needs.

Training Courses Consulting Services Training

PQC Protection for AI Model Weights and Training Infrastructure

Course Overview

Curriculum

Designed and Delivered By

Quantum Security Defence

Enquire About This Course