Ulascan Aytolun is experienced back-end developer, data scientist and security researcher. He worked as back-end developer at several companies to incorporate user needs into cost-effective, secure and user-friendly solutions known for scalability and durability. Also he worked as data scientist and has researches about some topics like fraud detection, suspicious behaviour detection and malware detection by using machine learning.
Making of ThreatScore: ML for Malware Detection
Three years ago, I wrote a graduation thesis about ML for Malware Detection. That work first lead to a fraud detection system for a global game company which has millions of daily users and then lead to Trapmine’s Threat Score Machine Learning Engine. Trapmine’s ThreatScore is a machine learning-based malware detection engine used by our customers and now scaling to be included in VirusTotal, scanning millions of files per day for viruses. The road from a inexperienced graduation thesis to a real product is long, compelling and took many engineers and security experts. This talk describes this journey and highlights the many challenges of delivering a malware classification product.