Data Scientist

Job Description

Location: Houston, TX

Type: Full-Time

Department: Data Science & AI

Reports to: Head of Data Science / Chief Data Officer

The Role

We re hiring a Data Scientist who turns data into revenue, efficiency, and innovation. You ll build predictive models, design experiments, and deploy ML solutions that solve real business problems from demand forecasting to personalization engines. If you live in Python, breathe statistical rigor, and ship models to production, this is your playground.

What You ll Do

Core Responsibilities

• Frame and solve high-impact problems with ML: churn prediction, pricing optimization, anomaly detection, recommendation systems.
• Design and analyze experiments: A/B tests, multi-armed bandits, causal inference, power calculations.
• Build, validate, and deploy models: Scikit-learn, XGBoost, LightGBM, PyTorch, TensorFlow from notebooks to APIs.
• Feature engineering at scale: SQL Spark, real-time streams, embedding layers.
• Own the ML lifecycle: Data prep training evaluation monitoring retraining.
• Collaborate with engineers: Model serving (SageMaker, MLflow, BentoML), CI/CD for ML, feature stores.
• Communicate insights: Storytelling with data, exec dashboards, model interpretability (SHAP, LIME).
• Stay ahead: Research papers, PoCs with LLMs, diffusion models, or graph neural nets.

What You Bring

Must-Have

• 3 years in data science or ML engineering.
• Python mastery: pandas, NumPy, scikit-learn, Jupyter, MLflow.
• Statistical depth: Hypothesis testing, confidence intervals, Bayesian methods, survival analysis.
• SQL fluency: Complex analytical queries, window functions, CTEs.
• Model deployment: REST APIs, Docker, cloud (AWS SageMaker, Google Cloud Platform Vertex, Azure ML).
• Experimentation rigor: You ve shipped A/B tests that moved KPIs.
• M.S./Ph.D. in Statistics, Computer Science, Math, Physics, or equivalent experience.

Nice-to-Have

• Deep learning: PyTorch/TensorFlow, transformers, computer vision, NLP.
• Big data tools: Spark, Databricks, Dask, Snowflake.
• MLOps: Kubeflow, Airflow, feature stores (Feast), monitoring (Evidently, WhyLabs).
• Causal inference: DoWhy, EconML, propensity scoring.
• Time-series forecasting: Prophet, ARIMA, LSTM, Temporal Fusion Transformers.
• Publications, Kaggle Grandmaster, or open-source ML contributions.

Tech Stack

• Languages: Python (primary), R (optional), SQL
• ML Frameworks: Scikit-learn, XGBoost, LightGBM, PyTorch, Hugging Face
• Data: Snowflake, BigQuery, PostgreSQL, Spark
• Deployment: SageMaker, MLflow, FastAPI, Docker, Kubernetes
• Experimentation: Eppo, Optimizely, Statsig
• Visualization: Streamlit, Plotly, Superset, Looker
• MLOps: Airflow, GitHub Actions, Terraform