Norfolk Southern is looking to solve unique business problems using novel techniques, such as detecting objects of interest or issues and defects using Deep Learning and Computer Vision algorithms, and predicting component failures using predictive models.
Requirements
- Scripting and programing experience with Python.
- Hands-on and theoretical knowledge of various machine learning and deep learning algorithm and frameworks such as: xgboost/LightGBM, Random Forests, SVMs, PCA, t-sne, kmeans, DBSCAN.
- Knowledge of Spark (PySpark) is a plus
- Knowledge of the cloud computing environment (e.g. Databricks) is a plus
- Expertise with Time Series problems.
- Familiarity with the railroad industry is highly preferred.
- Experience with Jupyter notebook, PyCharm, VS Code as IDEs.
Responsibilities
- Effectively utilize appropriate statistical, machine learning, and deep learning techniques to solve various business problems.
- Collaborate with various departments to identify opportunities for process improvement and developing analytics use-cases.
- Provide guidance, support and mentoring to junior team members.
- Evaluate accuracy and quality of data sources, as well as the designed models.
- Stays up to date with the latest models and changes in the technology.
- Design and develop (almost) production ready code.
- Communicate results to colleagues and business partners.
Other
- Masters degree in computer science, electrical engineering, machine learning, statistics, or related field.
- 3-5 years of experience as a Data Scientist, Research Scientist, Machine Learning Engineer or Operations Research.
- Advanced degree, e.g., Ph.D. or M.S. in deep learning, neural networks, machine learning, data science, computer science, electrical engineering, engineering, statistics, engineering, industrial systems, mathematical sciences, applied mathematics, or Physics.
- 5+ years of experience as a Data Scientist, Research Scientist, Machine Learning Engineer, or Operations Research.
- Hybrid 3 days on-site and remote work 2 days per week.
