AI Engineer (Contract)

Join a project focused on developing the "Risk Verdicator," an advanced LSTM-based engine designed to correlate security events into MITRE-mapped risks for an XDR platform. The role centers on building robust, explainable AI models capable of real-time inference (sub-3 seconds) using ONNX, with the goal of automating threat detection and response. You will work with mock datasets to train and deploy sequence models for anomaly detection, such as analyzing process trees and user behaviors. The position also involves implementing event correlation logic with CEP frameworks (Flink/Beam stubs), integrating SHAP for model explainability in the UI, and establishing MLflow for model versioning and Evidently for drift monitoring. The system aims for high accuracy (90% recall on MITRE ATT&CK corpus) and minimal false positives (<0.5%). Additional responsibilities include enriching the engine with intelligence mockups, outputting verdicts to GraphQL, and validating performance against industry benchmarks. Responsibilities Design, train, and deploy LSTM-based sequence models for anomaly detection using PyTorch and ONNX Implement event correlation using CEP frameworks (Flink/Beam stubs) Integrate SHAP for model explainability and UI feedback Set up MLflow for model versioning and Evidently for drift monitoring Enrich the engine with intelligence mockups and ensure robust output to GraphQL endpoints Validate models on the MITRE ATT&CK corpus, targeting 90% recall and <0.5% false positives Collaborate with stakeholders to refine requirements and deliver explainable, production-ready AI features Skills And Requirements • 3+ years of experience in Python and machine learning, with a strong portfolio in anomaly detection models • Proficiency in PyTorch, ONNX, and MLOps tools (MLflow, Evidently, Pandas/Beam pipelines) • Experience with event correlation, behavioral threat detection, or federated learning is a plus • Familiarity with SHAP, CEP frameworks, and GraphQL integration • Strong communication skills and ability to document and explain complex AI systems • Availability for part-time, contract-based work; remote candidates welcome About Twine Twine is a leading freelance marketplace connecting top freelancers, consultants, and contractors with companies needing creative and tech expertise. Trusted by Fortune 500 companies and innovative startups alike, Twine enables companies to scale their teams globally. Our Mission Twine's mission is to empower creators and businesses to thrive in an AI-driven, freelance-first world.