ML Engineer - Remote

About the Company

iAdeptive Technologies is an 8(a) small business providing data, cloud, and AI engineering services to federal mission programs. The team focuses on building reliable, secure, and scalable systems that meet demanding technical, security, and governance requirements. Engineers work closely with architects and other technical specialists, with an emphasis on practical implementation and accountable results.

About the Role

iAdeptive Technologies is seeking a Machine Learning Engineer to lead model development for a federal data and AI program. This is a hands-on modeling position focused on self-supervised learning, temporal data modeling, language-model-based information extraction, supervised classification, and model explainability.

You will own key modeling decisions, from architecture selection and training through calibration, validation, and interpretation of results. The work involves large, linked administrative datasets in a restricted analytical environment with approved tooling, no external connectivity, and CPU-only compute.

This role is best suited to an engineer with strong mathematical foundations who can evaluate competing approaches, explain modeling decisions, and deliver reproducible results. Data engineering supports the work, but model development and quantitative evaluation are the primary responsibilities.

Key Responsibilities

• Design and train self-supervised tabular transformers, including SAINT, TabTransformer, and FT-Transformer, along with temporal sequence models for longitudinal event data.
• Select pretraining objectives, run initial experiments, evaluate embedding quality, and benchmark candidate architectures against gradient-boosted baselines.
• Deploy locally hosted, open-weight language models to extract structured information from unstructured records using prompt design, constrained output schemas, and post-processing.
• Optimize models through quantization and CPU-inference tuning while maintaining acceptable extraction accuracy.
• Establish extraction quality through stratified sampling, independent human coding, field-level agreement measurements, and documented accuracy evaluations before large-scale processing.
• Train LightGBM, XGBoost, or CatBoost classifiers on learned embeddings and address severe class imbalance.
• Calibrate classification thresholds using validation data, accounting for the cost of different error types and reporting calibration curves alongside discrimination metrics.
• Identify and prevent data leakage across temporal splits and linked record sources.
• Use SHAP and other appropriate explainability methods to generate individual-level feature attributions and analyze patterns across cohorts and subgroups.
• Assess attribution stability across model retraining and confirm that explanations align with relevant domain expectations.
• Maintain supporting feature and inference pipelines using Python, Spark, and Parquet.
• Track datasets, models, and experiments in MLflow so reported results can be reproduced from recorded runs.

Required Qualifications

• At least 3 years of experience training and deploying machine learning models, including end-to-end ownership of at least one system from raw data through validated results.
• Hands-on experience training transformer architectures in PyTorch, including at least one self-supervised or contrastive learning objective applied to non-text data.
• Experience deploying open-weight language models for structured information extraction and evaluating accuracy against a labeled reference dataset.
• Practical knowledge of model calibration, class imbalance, and leakage prevention when working with real observational data.
• Experience with gradient-boosted methods such as LightGBM, XGBoost, or CatBoost and SHAP-based model explainability.
• Advanced Python skills, including NumPy, pandas, scikit-learn, and PyTorch, along with strong SQL skills.
• Experience using Spark for large-scale feature engineering.
• Strong applied understanding of probability and statistical inference, linear algebra, optimization, information theory, model evaluation, and causal reasoning.
• Ability to explain why modeling methods work, justify architecture choices against benchmarks, and make sound decisions in constrained computing environments.
• Eligibility to work in the United States and ability to obtain a Public Trust or higher background determination.

Preferred Qualifications

• A master's degree in mathematics, statistics, or applied mathematics with a concentration in data science is preferred. Computer science and other quantitative degrees may also be considered when academic transcripts and work experience demonstrate equivalent mathematical depth.
• Published or open-source work involving tabular deep learning, sequence modeling over event data, or language-model evaluation.
• Experience with model quantization and CPU-inference optimization at scale.
• Familiarity with healthcare claims or other administrative datasets, including diagnosis and procedure coding systems.
• Experience with fairness assessments and subgroup evaluation in applied machine learning.
• Experience working in FedRAMP, HIPAA, or other access-restricted environments.
• An active or previous federal Public Trust determination.

Technical Environment

• Tabular and temporal modeling: SAINT, TabTransformer, FT-Transformer, transformer encoders, and temporal convolutional networks.
• Language models: Locally hosted Llama, Mistral, or Qwen-family models, with quantized CPU inference using llama.cpp or equivalent tools.
• Supervised learning: LightGBM, XGBoost, CatBoost, and scikit-learn.
• Deep learning: PyTorch.
• Explainability: SHAP, including TreeSHAP and KernelSHAP, and integrated gradients.
• Data and experiment management: Python, Spark, Parquet, pandas, NumPy, and MLflow.
• Infrastructure: Cloud-hosted, access-restricted analytical environments.

What Success Looks Like

You will make evidence-based architecture decisions, build and evaluate self-supervised models, validate extraction quality before processing at scale, and establish classification thresholds based on measurable error costs. Model explanations should be meaningful and stable, while every reported result should be reproducible through documented experiments. The role also requires working effectively within approved libraries and CPU-only inference constraints.

Employment Details

• Location: Remote within the United States, with Maryland-area candidates preferred for occasional on-site collaboration.
• Employment Type: Full-time, W-2.
• Experience: 3+ years of machine learning model development and deployment.
• Education: A relevant master's degree is preferred, with equivalent quantitative depth considered.

Salary and Benefits

• Salary Range: $90,000–$120,000. The posted range reflects the floor and target for the position. Actual offers depend on experience, certifications, and clearance level.

• Benefits include health, dental, and vision coverage; a 401(k) with company contribution; paid time off and federal holidays; and training and certification support.

Equal Employment Opportunity

iAdeptive Technologies is an Equal Opportunity Employer. All qualified applicants will receive consideration without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, disability, protected veteran status, or any other characteristic protected by law.