Machine Learning Engineer, Fraud & Abuse Detection

At Telnyx, a leader in global connectivity solutions, the Machine Learning Engineer role is part of the Messaging team focused on fraud and abuse detection. The company is innovating in private multi-cloud IP networks and edge technology, aiming to modernize communication infrastructure. This role contributes to building real-time systems that filter millions of outbound messages daily to prevent phishing, smishing, brand impersonation, and other scams, supporting a secure messaging platform.

The engineer will design, train, and operate text classification models end-to-end, including data curation, labeling, evaluation, deployment, and monitoring for production drift. They will conduct shadow-mode evaluations to ensure model accuracy before enforcement, develop adversarial-resistant detection methods, and maintain strict latency requirements for inline message processing. Collaboration with fraud operations and engineering teams is essential, alongside participation in code reviews to uphold quality standards.

Candidates should have over three years of experience in software development or machine learning, with strong skills in asynchronous Python frameworks like FastAPI and production experience in text classification for fraud or abuse detection. Familiarity with large language models, rigorous evaluation techniques, and deploying ML models in production is required. Knowledge of messaging abuse types, adversarial ML, high-scale data systems, and cloud-native technologies is advantageous. The role is based in Dublin, São Paulo, or Buenos Aires, supporting a global, cross-functional environment.