ML Data Scientist (Models) - SageMaker & Python 12-Wk

JOHNSON SERVICE GROUP; Houston (TX); On-site; USD 138,000 - 158,000; Part time

Job summary

Johnson Service Group (JSG) is seeking a Data Scientist (Models) in Houston to develop predictive models for BOM/configuration mix and net manufacturing demand. This on-site, 12-week contract involves feature engineering and model experimentation alongside a Lead Data Scientist.

You will build classifiers, develop parallel model tracks, and document models with model cards and experiment logs to ensure traceability and accuracy reports.

Qualifications

• ML modeling experience across classification and regression.
• Strong Python skills with scikit-learn and LightGBM/XGBoost.
• AWS SageMaker experience, including training jobs and hyperparameter tuning.
• Ability to work under the direction of a Lead Data Scientist while independently supporting separate model development tracks.
• Strong technical documentation practices, including model cards, experiment logs, and accuracy reports.

Responsibilities

• Build a BOM/configuration mix classifier to predict series and part-number mix from family-group-level demand.
• Develop models in parallel with the Lead Data Scientist across separate model tracks to avoid single-threaded development.
• Develop a net demand model incorporating salvage, rerun, and repair signals to calculate total manufacturing need.
• Implement feature engineering pipelines designed by the Lead Data Scientist.
• Conduct model experimentation, including hyperparameter tuning, ablation studies, and feature importance analysis.
• Create and maintain model cards, experiment logs, accuracy reports, and related technical documentation.

Skills

Machine learning

Python (scikit-learn)

Python (LightGBM/XGBoost)

AWS SageMaker

Independent modeling

Model documentation

Job description

Johnson Service Group (JSG) is seeking a Data Scientist (Models) in Houston to develop predictive models for BOM/configuration mix and net manufacturing demand. This on-site, 12-week contract involves feature engineering and model experimentation alongside a Lead Data Scientist.

You will build classifiers, develop parallel model tracks, and document models with model cards and experiment logs to ensure traceability and accuracy reports.