Felix Waschke

PhD Student · Department of Software Science
Tallinn University of Technology (TalTech), Estonia
Freelance Detection Engineer

felix.waschke{at}taltech.ee

Research

My doctoral research is titled Using Machine Learning for Supporting Cyber Threat Handling in Security Operation Centers. Security Operations Centres (SOCs) receive far more alerts than analysts can investigate. I study how machine learning can help them prioritise and triage those alerts and detect threats. The methods have to work under real SOC constraints: little labelled data, heavy class imbalance and the need for analyst trust.

Consulting

Freelance Detection Engineering. I help security teams get detections that fire on real attacks, not on noise. I design, implement and tune custom SIEM rulesets, mapped to MITRE ATT&CK.

This builds on over five years of hands-on experience as a Detection Engineer and SOC Analyst, and on my PhD research on SOC alert handling. See my ML4SOC review (ACM Computing Surveys) and work on alert triage in SOCs (Computer Networks).

Ruleset audit Review your existing rules for coverage gaps against ATT&CK, dead rules and duplicates.
Custom detections New rules written for your environment, log sources and threat model.
Alert tuning Cut false positives on noisy rules so analysts can focus on real incidents.
Detection-as-code Version-controlled, tested rules with Sigma and CI deployment to your SIEM.

Platforms

Enquiries: felix@waschk.ee

Publications

2026

  1. Semi-supervised learning in Security Operations Centers: Performance, trade-offs and practical implications F. Waschke, A. Guerra-Manzanares, R. Vaarandi Computer Networks, vol. 289, 112697, 2026
  2. ML4SOC: A Comprehensive Review on Machine Learning for Security Operations Centres F. Waschke, R. Vaarandi, A. Guerra-Manzanares ACM Computing Surveys, vol. 58, no. 14, pp. 1–48, 2026

Teaching

Computer Forensics Lab → Interactive case studies for students: disk and file-system forensics in the browser, a network and SIEM investigation, and the hands-on "Learn" modules. AI Attack Lab → Hands-on LLM security: attack an AI assistant across scenarios on prompt injection, indirect and image injection, data exfiltration, and encoding bypasses, mapped to MITRE ATLAS.