Machine Learning Scientist

The Systems Optimization and Decision Analytics (SODA) Team is seeking a curious, innovative, and results-driven Machine Learning Scientist to help advance our AI and predictive modeling capabilities. We focus on building scalable, intelligent systems that power optimization, planning, forecasting, and human-in-the-loop decision-making for global operations. This role will center around designing and deploying cutting-edge ML models—including probabilistic models, large language models (LLMs), and time-series or agent-based systems. The ideal candidate brings deep expertise in modern ML techniques and a passion for turning theoretical innovation into production-ready systems that drive real-world impact.  

What You’ll Do: 

• Research, prototype, and implement state-of-the-art ML models across a range of tasks: forecasting, optimization, planning, recommendation, and human-AI teaming.

Develop models using advanced methods such as: 

• Large Language Models (LLMs), foundation model fine-tuning, and prompt engineering. 
• Probabilistic modeling, Bayesian inference, and uncertainty-aware decision systems. 
• Reinforcement learning (RL), multi-agent systems, and decision intelligence architectures. 
• Generative modeling (e.g., diffusion models, VAEs, normalizing flows). 
• Time-series and forecasting models (e.g., Temporal Fusion Transformers, DeepAR, N-BEATS). 
• Graph neural networks (GNNs), especially for spatio-temporal and structured prediction tasks. 
• Causal inference, self-supervised learning, and contrastive representation learning. 

Design and evaluate retrieval-augmented generation (RAG) and agentic workflows using LLMs. 

Scale experimentation and model training pipelines using Databricks, MLflow, and Spark. 

Partner with domain experts to frame complex, real-world challenges into solvable ML problems. 

Produce clean, reproducible code with strong documentation and CI/CD integration. 

What Skills You Need: 

• MS or PhD in Computer Science, ML, Statistics, or a related field. 
• Experience developing and deploying modern ML systems in production settings. 
• Solid foundation in deep learning and probabilistic machine learning. 
• Hands-on experience with transformer-based architectures, LLMs, and adaptation methods (e.g., fine-tuning, LoRA, RAG). 
• Strong Python skills with experience in PyTorch, TensorFlow, scikit-learn, and Hugging Face. 
• Familiarity with Databricks, Spark, and distributed computing frameworks. 
• Understanding of model evaluation, uncertainty quantification, and scientific experiment design. 

Nice To Have: