ML Engineer

━ About the Role We are looking for an ML Engineer to own model development on a federal data and AI program: self-supervised representation learning over tabular and temporal administrative data, language-model-based information extraction from unstructured records, supervised classification on learned embeddings, and model explainability. The work is performed against large linked datasets inside an access-restricted analytic environment with approved tooling, no external connectivity, and CPU-only compute. This is a modeling role, not a pipeline role. You will be accountable for architecture selection, training, calibration and validation of the models the program depends on, and for the quantitative judgment behind each of those decisions. Data engineering is part of the job; it is not the center of it. ━ What You Will Do Representation Learning Design and train self-supervised tabular transformers (SAINT, TabTransformer, FT-Transformer) and temporal sequence models over longitudinal event data, with the outcome label withheld during pretraining. Select pretraining objectives, execute pilot runs, and validate embedding quality before committing to full-scale training. Benchmark candidate architectures against gradient-boosted baselines and document comparative results. Language Model Extraction Deploy locally hosted open-weight language models for structured information extraction from unstructured text, including prompt design, constrained output schemas and post-processing. Quantize and tune models for CPU-only inference at volume without unacceptable degradation in extraction accuracy. Implement the dual-coding evaluation protocol — stratified sampling, independent human coding, field-level agreement measurement — and report accuracy before volume extraction. Supervised Modeling and Calibration Train gradient-boosted classifiers (LightGBM, XGBoost) on learned embeddings under severe class imbalance. Perform threshold calibration on the validation partition with the cost of each error type made explicit, and report calibration curves alongside discrimination metrics. Detect and prevent leakage across temporal splits and across linked record sources. Explainability Generate additive feature attributions (SHAP) at the individual-prediction level and aggregate them across cohorts and subgroups. Verify that attributions are stable across model retrains and consistent with domain expectations. Supporting Data Work Maintain the feature and inference pipelines the models depend on (Spark, Parquet, Python), and version every dataset, model and experiment in MLflow so reported results regenerate. ━ Mathematical Foundation This is a modeling position and we weight mathematical grounding heavily. Applied fluency is expected in the following; the interview will test it. Area Expected depth Probability and statistical inference: Estimators, bias and variance, sampling distributions; correct interpretation of confidence intervals and calibration curves Linear algebra and optimization: Gradient-based optimization, regularization, convexity and its failure in deep networks; attention understood as matrix operations Information theory: Cross-entropy, KL divergence, and the loss functions behind self-supervised and contrastive objectives Evaluation theory: Discrimination versus calibration; ROC and precision-recall behavior under class imbalance; threshold selection by expected cost Causal reasoning: Confounding and leakage; recognizing when a feature encodes the outcome or a downstream consequence of it ━ Technical Stack Layer Tools Tabular representation learning: SAINT, TabTransformer, FT-Transformer; self-supervised pretraining objectives (masked-feature reconstruction, contrastive) Temporal and sequence modeling: Transformer encoders over ordered event sequences (BEHRT-style), temporal convolutional networks Language models: Open-weight instruction-tuned models (Llama, Mistral, Qwen families) hosted locally; quantized CPU inference via llama.cpp or equivalent Supervised learning: LightGBM, XGBoost, CatBoost; scikit-learn Deep learning framework: PyTorch Explainability: SHAP (TreeSHAP, KernelSHAP), integrated gradients Data and MLOps: Spark, Parquet, pandas, NumPy; MLflow for experiment tracking and model versioning Environment: Cloud-hosted, access-restricted analytic environments ━ What We're Looking For 3+ years training and deploying machine learning models, with end-to-end ownership of at least one system from raw data to validated result. Hands-on experience training transformer architectures in PyTorch, including at least one self-supervised or contrastive objective on non-text data. Experience deploying open-weight language models for structured extraction, with accuracy measured against a labeled reference set. Demonstrated competence in calibration, class imbalance and leakage prevention on real observational data. Experience with gradient-boosted methods (LightGBM, XGBoost or CatBoost) and with SHAP-based explainability. Expert Python with NumPy, pandas, scikit-learn and PyTorch; strong SQL; Spark for large-scale feature engineering. Master's degree in mathematics, statistics or applied mathematics with a data science concentration is preferred. Computer science or other quantitative degrees are considered where the transcript and work record demonstrate equivalent mathematical depth. ━ Bonus Points Published or open-source work on tabular deep learning, sequence modeling over event data, or LLM evaluation. Model quantization and CPU-inference optimization at scale. Healthcare claims or other administrative data, including diagnosis and procedure coding systems. Fairness and subgroup evaluation methods in applied settings. Experience in FedRAMP, HIPAA or other access-restricted environments; active or prior federal Public Trust determination. Core Competencies Competency: What It Looks Like Here Architecture Judgment: Selects and justifies model architecture against a benchmarked baseline. Self-Supervised Modeling: Pretrains tabular and temporal transformers with the label withheld. Extraction Validation: Measures model-versus-human agreement before extracting at volume. Calibration Discipline: Reports calibration with discrimination and sets thresholds by cost. Mathematical Grounding: Explains why a method works, not only how to invoke it. Constrained Engineering: Delivers CPU-only inference inside an approved-library environment. Reproducibility: Every reported result regenerates from a recorded MLflow run. Why iAdeptive iAdeptive Technologies is an 8(a) small business that delivers modern data, cloud and AI engineering to federal mission programs. We are engineers first. We win work by building systems that hold up under audit and scale under real load — not by selling slideware. You will work alongside architects and engineers who write the code they design and treat security and governance as part of the build. Small enough that your work is visible; serious enough that it matters. Details Location: Remote (U.S.); Maryland-area candidates preferred for occasional on-site collaboration. Employment Type: Full-time, W-2. Eligibility to work in the U.S. required; this role supports federal programs and requires the ability to obtain a Public Trust or higher background determination. Education: Master's degree in mathematics, statistics or applied mathematics with a data science concentration preferred; computer science or related quantitative degrees considered with demonstrated mathematical depth. Experience: 3+ years training and deploying machine learning models with end-to-end system ownership. Benefits: Health, dental and vision coverage; 401(k) with company contribution; paid time off and federal holidays; training and certification support. Salary Range$90,000 – $120,000 Posted salary range reflects the floor and target for this role. Actual offers depend on experience, certifications and clearance level. iAdeptive Technologies is an Equal Opportunity Employer. All qualified applicants will receive consideration for employment without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, disability, protected veteran status, or any other characteristic protected by law.