Doctoral Seminar
Cyber Analytics
Security fails when it is treated as a list of findings. Twelve chapters on treating it instead as measurement, inference, prediction, validation, simulation, and governed action — each claim carrying the evidence that licenses it.
The course surveys analytical tools for cybersecurity data and emphasizes how data analytics procedures support defensible cybersecurity policy decisions across discovery, diagnostics, vulnerability assessment, risk modeling, simulation, cryptographic readiness, federated intelligence, supply-chain integrity, deception, formal verification, and closed-loop remediation.
Instructor
Instructor
Taught by Dr. Ravi Mallarapu, who brings ~30 years of industry experience spanning algorithmic trading, cloud security, and cyber defense — including senior roles at Bridgewater Associates and D. E. Shaw, and current work leading AI-driven systematic trading research at Blockchain.com and advising NATO's Collective Cyber Defense Center.
Full bio →Inside the Book
One rigorous volume follows the complete operational security pipeline, from measurement through governed action.
Network Discovery and Asset Inventory
What a scan can and cannot establish. Vantage, coverage, and the discovery gap between the hosts that exist and the hosts you admitted.
Service Enrichment and Device Fingerprinting
Turning banners, certificates and protocol behavior into identity claims, and the confidence ceiling each surface can support on its own.
Vulnerability Assessment
Matching software to advisories when names disagree, and reporting what the evidence licenses instead of a single confident severity number.
Attack Graph Analytics and Contextual Risk
Attack graphs as an ordering over paths rather than a probability, with the margin at which a ranking reverses reported alongside the ranking.
Autonomous Penetration Testing under a Safety Lattice
Testing an exposure claim by acting on it, bounded by an authorization model, a blast radius, and an abort path exercised before the run.
Digital Twins for Remediation Simulation
Fidelity as a vector rather than a score, and the monotonicity argument that decides when a twin's verdict may be acted on.
Post-Quantum Cryptographic Readiness
Post-quantum readiness as decision analysis over scenario sets, where the binding constraint is migration effort, not the arrival date of a quantum computer.
Federated Threat Intelligence
Federated inference across organizations, with exact differential-privacy accounting and a robustness bound that is honest about what it does not cover.
Supply-Chain Integrity and Counterfeit Detection
What a bill of materials establishes about a component, what it cannot establish about the hardware, and completeness as a metric with a denominator.
Active Deception and Threat Hunting
Deception as an instrument that produces evidence, its yield modeled explicitly, and the attribution claims a source address cannot carry.
Formal Protocol Verification
Symbolic protocol analysis under a Dolev-Yao adversary, and the discipline of distinguishing proved from tested from assumed.
Autonomous Remediation and Safety Verification
Remediation as an ordered plan under explicit authorization, where rank reversal across objectives is reported rather than resolved by a weighting.
Machine Learning Masterclass
Learn supervised learning, metrics, calibration, boosted trees, causal sequence models, replay, and governed automation through one coherent research story.