Vikta Energy

Senior Data Scientist

US$160,000-$220,000Posted 2 days ago

RemoteFull-Time Machine Learning Data Science $200k+

Job Description

ABOUT VIKTA ENERGY

Vikta Energy is building the world’s leading intelligent energy platform and the largest independent Tesla Supercharger network in North America. Vikta combines Tesla’s industry-leading charging technology with proprietary AI, machine learning, and software to identify the highest-value locations, maximize utilization, and connect charging customers with the businesses where they charge, creating new value for drivers, retailers, and property partners.

ABOUT THE ROLE

Reporting directly to the SVP of Data Science and Decision Intelligence, you’ll be one of the first members of the data science team. You’ll work on building Vikta's data science function from the ground up. You'll own modeling, pipelines, and analytics end-to-end, working on key site-selection and utilization decisions. This is a high-autonomy, high-ambiguity role suited to someone who's comfortable owning a problem from messy data through to a production prediction and a stakeholder-ready insight.

WHAT YOU’LL DO

ML modeling. Develop, validate, and maintain ML models, starting with the EV charger utilization model and expanding to others as the roadmap grows.
Feature engineering. Source, clean, and engineer features from various datasets.
Data pipelines. Build and maintain reproducible ingestion and transformation pipelines that feed both models and reporting, with attention to data quality and lineage.
Dashboards & BI. Build interactive dashboards and reporting that make model outputs and business metrics legible to non-technical stakeholders.
Analytics support. Serve as an analytics expert to the rest of the company. Provide ad hoc analysis, decision support, and turning open-ended business questions into tractable data work.

WHAT YOU’LL BRING

Required

6+ years building predictive models that were used in production. Ownership from problem framing through deployment, monitoring, and retirement.
Production experience with gradient-boosted trees (LightGBM, XGBoost). Can explain how they work, how to improve them, their limitations, etc.
Spatial data fluency. geopandas, PostGIS, H3 or equivalent. Understands spatial autocorrelation and why random k-fold leaks on geographically distributed targets.
Validation discipline. Designing holdouts that mirror deployment conditions, calibration, uncertainty quantification, and the judgment to know when a metric gain is noise.
Inference from observational data. Small samples, no randomization, confounded selection. Comfortable reasoning about bias and running sensitivity analysis. This is not an A/B testing role.
Strong Python (pandas, NumPy, scikit-learn) and SQL beyond basic joins. Window functions, query tuning against a warehouse.
Feature engineering from messy third-party data. Reconciling inconsistent geographies, identifiers, and vintages across sources.
Reproducible workflows. Git-based development, versioned data and experiments (DVC, MLflow, or equivalent).
Can derive, defend, and debug your work independently. AI tooling is expected and encouraged for speed, but you will be asked to justify modeling choices from first principles in review, and to diagnose why a model is wrong.
Raises the technical bar on people around you. Reviewing others' modeling work, setting conventions, making the team's output better.
Translates model output into decisions executives and site development teams act on.
Self-directed under ambiguity. Comfortable defining the problem and adjusting as the project continues.