Like the Internet, the Telecommunications networks were initially built without security considerations. Various vulnerabilities have been discovered across telecommunication protocols. Notable attacks are location tracking, spoofing, call and SMS interception, network stack exploitation, denial of service, and SMS Phishing. It is also crucial for operators to build a monitoring solution that ensures visibility and detects malicious activities. Attackers have increasingly used SMS Phishing as the cost of sending SMS is getting negligibly. Detecting SMS Phishing attacks in real-time while respecting user privacy is an important usecase for Telecom operators. In this talk, we will first give a high-level background introduction about Telecom threats. We will then focus on the SMS Phishing usecase and share our experience and technical details about using Machine Learning in building a real-time detection system from scratch.