As a Senior Data Scientist on Inspire’s Data Platforms and Services team, you will use modeling and statistical tools to define, deploy, maintain and improve the use of forecasting and algorithmic learning to drive business objectives and optimize operations. We take a product-driven, agile approach to our platform, driving measurable growth and meaningful outcomes every single sprint. We build efficient, scalable processes in a service-oriented ecosystem leveraging powerful code frameworks and repeatable patterns to solve real problems for stakeholders and customers.
THE X HAS 4 MAIN RESPONSIBILITIES
- Apply state-of-the-art machine learning and predictive modeling techniques to provide a deeper understanding of customers, and develop and improve the accuracy of our energy and gas forecasts
- Build processes by which we monitor performance of our forecasting services over time
- Develop, improve, and extend the implementation and architecture of our forecasting services
- Communicate modeling and engineering progress to stakeholders and partners to drive impact and facilitate decision making
SOME YEAR 1 DELIVERABLES
- Develop an energy load forecast service with interval meter data
- Develop a natural gas forecast service
- Improve the accuracy of our short and long-term energy load forecast services
- Improve the configurability and robustness of our forecast services
- Implement input and output data monitoring of our load forecasts
- Conducted experiments to demonstrate forecast development and iterative improvement
- Communicated improvements to stakeholders
- Developed monitoring solutions for input data and forecast outputs
- Developed technical project plans and delivered results
- Technical competency - comfort on a command line, a good grasp on the fundamentals of programming, familiarity with Git/source control, in-depth knowledge of domain-specific tools and frameworks
- Statistical competency - able to navigate and apply statistical frameworks for measuring confidence and predictive power. Aware of how assumptions may be violated or models over-fit, able to caveat findings with technical limitations and common sense.
- Results-orientation - resists the urge to get caught up in a great idea, emphasizes testing, and aims for measured outcomes that align with business value.
- Problem-solving mentality - gets excited about digging into complexity, wants to ask questions and learn more, and isn’t put off by problems they’ve never been explicitly told how to solve. Especially troubleshooting: ability to break down a chain of steps to narrow and locate a problem.
- Big-picture awareness - Understanding of the importance of context, and ability/willingness to understand the business problem in addition to the technical one. Focus on people & impact. Identify shortcuts & justify appropriate level-of-effort. Pre-emptive identification of potential issues downstream.
- Must Have
- 5-8 years of experience applying machine learning and statistical knowledge to drive business value
- 5+ years of Python experience
- 5+ years of experience using SQL to query large datasets
- Excellent communication skills, with the ability to deliver complicated findings and explain technical approaches to a variety of audiences
- Fluency with data visualization to communicate complex topics in approachable ways
- Nice to Have
- Experience developing energy load forecasting models using interval meter data, preferably in ERCOT
- Experience developing natural gas forecasting models
- Experience with data engineering frameworks: Apache Airflow, dbt, MLFlow, Apache Spark, AWS services, Docker, Kubernetes
- Experience with machine learning frameworks: scikit-learn, Light GBM, XGBoost, lifelines, pytorch, Tensorflow/Keras
- Software development lifecycle experience in GitHub (ie environment management, testing, deployment)
- Experience at a similar scale of data processing (Multi-TB/billions of rows)
- Work with real-time event stream data
- Contextual work in the energy industry