Job Description
Job Description
This is a remote position.
THIS IS A FULLY REMOTE POSITION - CANDIDATE MUST BE LOCATED IN THE U.S. (Visa sponsorship NOT available). This position will architect and evolve our client's decision and action engine that detects churn, optimizes pricing, predicts demand, and automates retention campaigns. This role develops the models, orchestration logic, and continuous-learning loop that make Command proactive and measurable.
Key Responsibilities
- Design, train, and deploy ML models for :
- Customer defection / retention scoring (LightGBM / XGBoost + survival models + SHAP explainability).
- Dynamic service pricing (gradient boosting + contextual bandits / Bayesian optimization).
- Appointment and capacity optimization (Google OR-Tools + LightGBM).
- Parts risk & first-time-fix forecasting (time-series + anomaly detection).
- Message generation using hosted GPT-class LLMs or Llama 3.x with retrieval-augmented prompts.
- Build model-training pipelines and monitoring dashboards integrated with Command’s backend.
- Develop AI orchestration layers (LangChain, CrewAI, or equivalent) that link model outputs to actions.
- Continuously retrain models as Command scales across rooftops to improve accuracy and ROI prediction.
- Partner with backend engineer to unify data ingestion → decision → action → feedback workflow.
Requirements
Python expertise with TensorFlow, PyTorch, scikit-learn.Proficient in ML lifecycle tools (MLflow, Weights & Biases).Experience with recommendation systems, churn prediction, and reinforcement learning for pricing.Strong data-engineering foundation (SQL, feature engineering, model validation).Understanding of LLM orchestration and retrieval-augmented generation for message automation.Familiarity with automotive or large-scale consumer service data preferred.