Venture Global LNG (“Venture Global”) is a long-term, low-cost provider of American-produced liquefied natural gas. The company’s Louisiana-based export projects service the global demand for North American natural gas and support the long-term development of clean and reliable North American energy supplies. Using reliable, proven technology in an innovative plant design configuration, Venture Global’s modular, mid-scale plant design will replace traditional designs as it allows for the same efficiency and operational reliability at significantly lower capital cost.
The Data Scientist is responsible for designing, developing, maintaining, and deploying machine learning, simulation, and optimization models using Databricks and Spark (PySpark). Strong data engineering, communications, and self-management skills are a must.
This position is for experienced candidates only. That means the candidate will have to demonstrate real world experience (not academia) as well as the expertise gained from that experience.
The position reports to the Director of Business Intelligence and is structured within IT under the Vice President of Applications.
The position is located in Arlington, VA and requires working at the office five days a week.
Responsibilities
- Build end-to-end data science workflows, not just models, that provide quantifiable business value to stakeholders.
- Work the whole data science lifecycle: requirements, data generation, data transformation, model building, model testing, model serving.
- Provide software solutions that are automated and flexible to future business demands. Mercilessly fight technical debt and rework.
- Develop and leverage subject matter expertise in select portions of the business in order to better design and execute projects.
Qualifications
- Bachelor’s degree in analytical field.
- 5 years experience data science work outside academia.
- 2 years experience in Spark (PySpark) OR experience in other Python data libraries plus some exceptional demonstration of data science ability. Writing SQL or Pandas code in Spark does not count.
- Excellent English.
- Excellent communication skills, with a history of being able to communicate in oral, written, and presentation form with senior stakeholders such as Directors, Vice Presidents, and the C-Suite.
- The ability to work in a completely self-directed manner. This includes the ability to learn complex software by one’s self.
- The ability to organize complex work in a formal way so as to be able to keep track of multiple complex projects for multiple stakeholders.
- Broad understanding of how to write and deploy good code. This means being able to write code in functions, not just notebooks. Other topics are Git / version control, design of modular code to separate out constants and lookup tables, unit testing, and writing readable textbook-style code that all team members can understand.
- Thorough understanding of basic data engineering concepts such as relational and nonrelational data stores, distributed computation, batch and stream processing, tables, keys, SQL vs Spark, cardinality, transforms (groupby, agg, melt, pivot, window, join, union, merge, etc), partitions, caching, and so on.
- Thorough understanding of at least one Python visualization library such as matplotlib, seaborn, plotly, holoviz, altair, etc. Excellent understanding of both business-centric visualization and scientific visualization, including how to best display quantitative information to nontechnical audiences. Ideally, a mature understanding of the aesthetics of visualization.
- Thorough understanding of applied data science in the form of machine learning. Knowledge of the statistical principles that underpin machine learning, the basic types of machine learning models, and the proper use of Python ML libraries such as scikit-learn and spark.ml / mllib.
Preferred Qualifications
- Knowledge of additional data science methodologies beyond vanilla machine learning: simulation methodologies such as system dynamics (SD), agent-based modeling (ABM), and discrete event simulation (DES); deep learning; the application of ML / DL to unstructured data such as text, audio, images, and video; and optimization.
- Demonstrated expertise in applying one or more data science methodologies in an advanced way to achieve business impact. Writing model.fit() and model.predict() is not a sufficient demonstration of data science talent.
- A sophisticated understanding of how to push past existing common practice in the application of data science as relevant to Venture Global LNG. The goal should be to radically, not incrementally, improve data science capabilities every year.
Venture Global LNG is an Equal Opportunity Employer. We do not discriminate on the basis of race, religion, color, sex, gender identity, sexual orientation, age, non-disqualifying physical or mental disability, national origin, veteran status, or any other basis covered by appropriate law.
#LI-Onsite
