Breaking the Mobile Log Analysis Barrier with AI
Numa Dhamani
BSides NYC 2025 (0x05) · Day 1
In the realm of cybersecurity, mobile device logs represent an invaluable, yet often inaccessible, trove of forensic data. Numa Dhamani, Head of Machine Learning at iVerify, delivered a compelling talk at BSides NYC, "Breaking the Mobile Log Analysis Barrier with AI," addressing this critical challenge. The presentation highlighted how the sheer scale, complexity, and fragmentation of mobile logs—such as Android bug reports and iOS unified logs—create a significant barrier to effective security investigations and incident response.
AI review
Dhamani describes a real operational problem — mobile logs are genuinely terrible to work with at scale — and the iVerify pipeline has some thoughtful engineering in it (PII redaction before LLM ingestion, semantic chunking instead of naive splits, evidence grounding for researchers). But the talk never escapes the product demo gravity well long enough to teach the audience something they couldn't get from reading the Bugalyzer landing page.