About the Role
You will design, train, and deploy machine learning models to detect anomalies, mitigate bots, and identify zero-day intrusions at the edge. You will implement efficient inference pipelines that process massive volumes of HTTP traffic with ultra-low latency. You will collaborate with offensive security engineers to turn attack data into high-quality training datasets, build and maintain data pipelines that ingest traffic logs, monitor model performance, detect concept drift, and automate retraining. You will write production-ready services and integrate models using Python, Rust, or TypeScript, deploy and operate models on cloud and edge platforms, and evaluate LLMs and agents where applicable. You are expected to work autonomously in a fully remote environment and communicate technical concepts in English and Portuguese.
Requirements
- Fluency in Python and SQL
- Proficiency with ML libraries such as PyTorch TensorFlow or Scikit-Learn
- Experience deploying machine learning models into production with high-throughput low-latency constraints
- Strong understanding of web security HTTP protocols and common attack vectors (OWASP Top 10 L7 DDoS credential stuffing)
- Familiarity with edge computing platforms such as Cloudflare Fastly or AWS Edge and CDN WAF concepts
- Solid software engineering fundamentals and ability to write production-ready services
- Experience with cloud infrastructure (GCP AWS) and managing large-scale datasets
- Experience with LLMs and agents
- Ability to work autonomously in a fully remote distributed team and communicate effectively in English and Portuguese
Responsibilities
- Design and train machine learning models for anomaly detection bot mitigation and zero-day intrusion detection
- Deploy and optimize inference models to run at the edge with ultra-low latency
- Process massive volumes of HTTP network traffic for real-time detection
- Collaborate with offensive security engineers to generate and label attack datasets
- Build and maintain data pipelines that ingest traffic logs and automate retraining
- Monitor model performance and detect concept drift
- Integrate models into production services using Python Rust or TypeScript