TekSynap is seeking an AI/ML Engineer to support the National Geospatial-Intelligence Agency (NGA) Artificial Intelligence (AI) efforts by creating technical solutions for multiple AI challenges within a secure, agentic AI framework.
Requirements
- 3+ years of experience designing, training, validating, and deploying ML models (especially computer vision, geospatial analytics, or time-series)
- Proficiency in Python, PyTorch/TensorFlow, and ML lifecycle tools (MLflow, Kubeflow, Airflow)
- Experience building data/model ingestion pipelines, metadata registries, and automated scoring/benchmarking frameworks
- Familiarity with containerized deployment (Docker/Kubernetes), REST APIs, and distributed architectures (e.g., AWS EventBridge).
- Working knowledge of model governance, explainable AI (XAI) methods, and regulatory/compliance constraints for AI in secure environments
- Experience implementing: Drift detection algorithms and active retraining loops
- Familiarity with Model Context Protocol (MCP), Agent-to-Agent (A2A) communications, and multi-agent orchestration frameworks
Responsibilities
- Design and implement model ingestion pipelines to onboard new models and datasets
- Automate extraction, transformation, and loading (ETL) of model metadata (hyperparameters, training datasets, evaluation metrics, licenses) into standardized registries for validation and governance
- Integrate metadata into unified data ontology structures to enable downstream discoverability and auditability
- Develop automated model evaluation loops to continuously test and benchmark models against mission-defined performance thresholds
- Build scoring logic and performance scorecards, integrating with agent monitoring frameworks to detect constraint violations and trigger re-validation workflows
- Implement continuous data/model drift monitoring and mitigation pipelines; create retraining or fallback workflows to maintain operational accuracy
- Build explainability layers (feature importance, confidence bounds, constraint violation predictions) and automated factsheets for model transparency and audit compliance
Other
- Collaborate with governance teams to align validation pipelines to IC/Department of Defense standards and National Geospatial-Intelligence Agency accreditation criteria.
- Enable human-in-the-loop checkpoints through interactive dashboards, allowing analysts to provide feedback and influence model lifecycle decisions
- Active TS security clearance, CI Poly eligible
- U.S. Citizenship
- TS/SCI with Polygraph
