Python AI Engineer (Risk Verdicator Feature)

You’ll build the “Risk Verdicator”—an LSTM-powered engine that correlates events into MITRE-mapped risks, with ONNX inference for latency under 3 seconds. This system will power our XDR platform, aiming for explainable, robust AI that can automatically trigger responses. You will train and deploy sequence models using PyTorch and ONNX for anomaly detection (like on process trees or user behaviors) on mock datasets. There will be a need to implement event correlation using CEP (Flink/Beam stubs) and SHAP to provide UI explanations, along with setting up MLflow versioning and Evidently for drift monitoring, targeting less than 0.5% false positives. Additional work includes enriching the engine with intelligence mockups, outputting verdicts to GraphQL, and validating on the MITRE ATT&CK corpus with a 90% recall goal. The ideal candidate has over 3 years of Python/ML experience, with a portfolio in anomaly models, experience in MLOps tools (MLflow, Evidently, Pandas/Beam pipelines), and strong communication skills. Experience with behavioral threat detection or federated learning is a plus.

Budget range:
Receive offers

Who are you, and what do you do?
Managing Director

In what capacity are you hiring?
For my company

Where are you in the hiring process?
I’m ready to make a paid hire

What type of work is this?
Part-time position (Contract)

What experience level is needed?:
Beginner