Description
nlp modeling
predictive modeling
clustering
ensemble modeling
model monitoring
data pipelines
Electric T&D Engineering is responsible for electric system engineering, planning, asset strategy, and risk management across transmission, distribution, and substation asset families. This centralized, risk‑informed approach enables consistent standards, work methods, prioritization, and program sponsorship while leveraging inspection lessons and asset data to guide asset management decisions.
- Develop NLP‑based and supervised learning models to automatically classify electric incidents as reportable or non‑reportable and identify safety incident types and severity.
- Build predictive models to assess the risk of CPUC Notice of Violation for new incidents, analyzing historical patterns by asset type, region, and reporting type.
- Apply unsupervised learning to cluster incident data, create archetype profiles, and map clusters to preventative programs and corrective actions.
- Use decision trees, random forests, and gradient boosting to recommend evidence‑based root‑cause evaluations.
- Implement continuous model monitoring to detect data drift, label shift, performance decay, and regulatory changes.
- Engineer robust data pipelines (ETL) and reusable Python code for feature engineering, collaborating with sponsor departments and subject‑matter experts to align models with business needs.
- Present findings and recommendations to senior leadership and serve as peer reviewer for complex models while actively contributing to the external AI/ML community through volunteering, conferences, or publications.
- Ensure adherence to data‑science standards and best practices for model evaluation, optimization, and deployment.
Requirements
bachelor's
phd
6+ years
python
ml
feature engineering
PG&E seeks an experienced data‑science professional to join the Electric Compliance Assurance, Analysis & Intake (A&I) Team as a Data Scientist, Expert. The role focuses on creating advanced machine‑learning and predictive‑modeling solutions to improve electric and safety incident classification, reduce CPUC late‑reporting violations, predict compliance issues, and speed root‑cause evaluations.
The position is headquartered at PG&E’s Oakland General Office and may require occasional travel across the company’s service territory.
- Require a bachelor’s degree in a quantitative field and at least six years of data‑science experience (or a doctoral degree in lieu of experience).
- Preferred qualifications include a doctorate, utility or renewable‑energy industry experience, advanced Python proficiency (pandas, scikit‑learn, NLP libraries), and demonstrated expertise in supervised/unsupervised learning, feature engineering, and communicating technical insights to non‑technical audiences.
Benefits
The salary range for this Bay Area position is $140,000 minimum, $189,000 mid, and $238,000 maximum, with the final offer based on factors such as skills, education, experience, geographic location, and internal equity.
Training + Development
Information not given or found
