Data Scientist

Job Title: Data Scientist

Job Category: Science

Time Type: Full time

Minimum Clearance Required to Start: TS/SCI

Employee Type: Regular

Percentage of Travel Required: Up to 10%

Type of Travel: Local

* * *

The Opportunity:

The 35th Intelligence Squadron seeks a motivated AI/ML Engineer to develop and deploy complex Artificial Intelligence systems defending the Department of Defense Air Force Information Network (AFIN). The role requires anomaly detection algorithms for identifying and isolating malicious threats using supervised and unsupervised learning, Graph Neural Networks, and Deep Learning models.

You will drive our cyber threat intelligence and detection mission by providing expert leadership and mentorship in advanced AI/ML solutions, augmenting cyber threat analysts triaging 1TB+ of boundary device logs daily to produce defensible intelligence reports with real mission consequences. You will provide technical direction for the design, implementation, testing, deployment, and operation of the 35 IS's cyber threat detection methods and enabling systems.

Responsibilities:

• Design & Deploy: Architect, build, and deploy high-performance ML models with 1TB+ daily ingestion across heterogeneous data sources into production, ensuring scalability, reliability, and low latency.

• Drive the ML Lifecycle: Lead data preparation, model development, evaluation, monitoring, drift detection, and continuous retraining within mission-aligned constraints.

• Model Innovation: Develop and implement state-of-the-art algorithms, specifically a hybrid GNN and BiLSTM architecture operating on a continuously updated heterogeneous network graph.

• Optimize: Improve model performance through feature engineering, hyperparameter tuning, and advanced experimentation.

• Mentorship: Mentor junior software developers and provide technical guidance and expertise.

• Logistics: Ensure appropriate documentation for all delivered analytics. Build explainability into the product from the beginning to support defensible intelligence reports. Ensure analyst trust is a design requirement, not an afterthought.

• Innovation: Guide analytic approaches where information is incomplete or no precedent exists. Apply that experience to real constraints: classification boundaries, data that cannot be shared with vendors, upstream pipeline failures, and out of order log delivery.