AI/ML Lead Engineer

Location - NJ/TX/AZ /RI Responsibilities • Design, prototype, and implement applied AI/ML models for mobility, activity, and behavioral modeling at scale. • Drive improvements to core models for routing, mode choice, demand estimation, and geospatial forecasting — balancing research innovation with production readiness. • Partner closely with Product and Data Science leadership to set technical direction for modeling initiatives, ensuring they align with customer and business needs. • Evaluate new data sources (e.g., vendor-provided mobility data, traffic sensors, geospatial layers) for quality, coverage, and fitness-for-use in Replica’s models. • Develop new algorithms to fuse Replica’s proprietary data with external sources (e.g., traffic sensors, land use, census, streaming data). • Evaluate model performance, define success metrics, and iterate quickly to improve accuracy, robustness, and fairness. • Collaborate with data production engineers to transition prototypes into robust, production-quality systems used nationwide. • Contribute to modeling frameworks and shared libraries, mentoring peers and raising the technical bar across the team. • Stay current with advances in machine learning, optimization, and geospatial modeling, applying them pragmatically to Replica’s challenges. Minimum Qualifications • Bachelor’s degree in Computer Science, Applied Mathematics, Engineering, or related field • 4+ years of professional experience in applied ML/AI, research engineering, or related roles. • Strong Python skills with experience in ML libraries (PyTorch, TensorFlow, scikit-learn) and numerical computing (numpy, pandas, scipy). • Experience designing, training, and deploying AI/ML models in production. • Hands-on experience with large-scale data sets, especially geospatial or time-series data. • Familiarity with distributed computation frameworks (Dask, Spark, Ray) and cloud environments (GCP preferred). • Proven ability to communicate technical decisions and tradeoffs clearly to both technical and non-technical stakeholders.