Senior / Staff Applied Machine Learning Engineer

We are hiring a Senior or Staff Applied Machine Learning Engineer to develop predictive capabilities over a temporal, geospatial knowledge base assembled from customer, public, and proprietary data. The role covers anomaly and pattern-of-life detection, risk estimation, forecasting, entity matching, confidence calibration, model evaluation, and production inference.

This is a hands-on individual contributor role on an existing product and engineering team. You will write production code regularly, develop training and inference systems end-to-end, and ship measured capabilities into operational environments. Early work will emphasize evaluation, data quality, and practical customer-backed models before expanding into broader representation-learning or world-modeling research. The position is on-site and requires U.S. citizenship and eligibility to work on ITAR/export-controlled programs.

• Develop predictive use cases with clear operational value, including anomaly and pattern-of-life detection, risk estimation, logistics or sustainment forecasting, trajectory analysis, and relationship prediction.
• Build statistical and learned representations for heterogeneous temporal, relational, and geospatial data while retaining provenance and uncertainty.
• Develop models that improve entity matching, extraction, evidence reconciliation, ranking, and confidence estimation.
• Establish deterministic and statistical baselines before introducing more complex modeling approaches.
• Build training and evaluation datasets from sparse, delayed, incomplete, or contested ground truth.
• Design experiments that account for temporal leakage, geographic leakage, source dependence, changing data distributions, and correlated failure modes.
• Evaluate and improve model calibration, including reliability analysis, confidence intervals, isotonic regression, Platt scaling, or other appropriate methods.
• Define performance measures such as false alert rate, prediction lead time, calibration error, intervention frequency, and performance under source loss.
• Build replay, shadow-evaluation, model-versioning, promotion, rollback, and production-monitoring pipelines.
• Incorporate analyst feedback, assessment corrections, and post-event outcomes without treating every interaction as verified ground truth.
• Maintain reproducible training lineage, dataset versions, model artifacts, evaluation results, and promotion decisions.
• Define degraded-mode behavior, fallback strategies, cached-output policies, and alert thresholds for model or source unavailability.
• Adapt models for cloud, on-premises, and disconnected environments using efficient inference, quantization, distillation, or alternative model classes where appropriate.
• At the Staff level, set the applied ML agenda, establish evaluation and serving standards, and mentor other engineers.

• U.S. citizenship, with eligibility to work under ITAR/export-control programs. Active or prior security clearance is a plus.
• Bachelor's or Master's degree in Computer Science, Statistics, Applied Mathematics, Machine Learning, or a related field, or equivalent practical experience.
• 5+ years building and operating production ML systems, with demonstrated ability to ship models into real operational environments.
• Strong Python engineering skills with production-grade software development practices.
• Experience across the ML lifecycle, including data preparation, training, evaluation, deployment, monitoring, and iteration.
• Depth in at least one relevant area: spatiotemporal modeling, forecasting, graph or relational learning, anomaly detection, information extraction, ranking, or entity matching.
• Solid understanding of experimental design, leakage prevention, probabilistic calibration, uncertainty, and error analysis.
• Experience working with sparse, delayed, incomplete, changing, or contested ground truth.
• Experience with modern ML frameworks and production data infrastructure.
• Ability to define a useful operational decision, identify available evidence and labels, build a credible baseline, and measure whether a model improves the outcome.
• Clear, concise written and verbal communication.

• Experience with Bayesian inference, survival analysis, probabilistic forecasting, or calibrated ensemble methods.
• Experience with graph representation learning, link prediction, or heterogeneous knowledge graphs.
• Experience with geospatial, trajectory, remote-sensing, or time-series modeling.
• Experience with vector databases, graph databases, PostgreSQL/pgvector, or lakehouse and columnar analytical storage.
• Experience operating NVIDIA Triton Inference Server or an equivalent production model-serving platform.
• Experience with efficient inference, quantization, distillation, or edge deployment.
• Experience building adversarial robustness tests for models operating on untrusted or potentially manipulated data.
• Familiarity with MCP or equivalent tool-layer orchestration patterns.
• TypeScript proficiency for integration with agent tools or product interfaces.
• Experience deploying ML systems in restricted, on-premises, or disconnected environments.
• Experience in defense, intelligence, public safety, or another high-consequence domain.