Part-Time ML Engineer / Applied Scientist - Internship

Summary: AXYS, Inc. is an early-stage company focused on building an enterprise ML prediction platform. They are seeking a part-time ML engineer/applied scientist to work on validating existing ML pipelines and processing real customer data, with a focus on predictive maintenance and model evaluation. Responsibilities: - Work through real customer data, SAP PM work orders, equipment event streams, sensor data, and operator-coded logs - Clean and process operational data: missing labels, censored observations, noisy categorical fields, time-alignment issues - Engineer features for sensor and event data across diverse equipment types. We build per-equipment-type models, not pooled across the line - Train and evaluate models across prediction, classification, anomaly detection, and survival analysis (Weibull AFT, Cox PH, competing risks, RUL) - Strengthen anomaly detection, currently the lightest area of our modeling architecture - Contribute to end-to-end ETL and training pipelines on our AWS stack - Pressure-test modeling and architectural decisions as the platform grows Required Qualifications: - Master's degree or above in CS, Statistics, ML, Applied Math, Engineering, or a related quantitative field. Current PhD candidates and postdocs are welcome - Hands-on applied ML experience, you have personally built and evaluated models end-to-end, ideally including at least one production deployment - Solid applied-ML methodology: feature engineering and evaluation on messy real-world data; awareness of label noise, leakage, drift, selection bias, and censoring - Proficient in Python and the standard applied-ML stack (pandas, scikit-learn, lifelines or equivalent, etc.) - Comfortable working in an AWS environment (SageMaker, Lambda, Step Functions, DynamoDB, S3) — or able to come up to speed quickly - US Residency/work permit is required Preferred Qualifications: - Experience with time-to-event / survival methods (Weibull AFT, Cox PH, competing risks, RUL) - Experience with operational, industrial, or sensor data: ERP / MES exports, transactional records, human-coded categorical fields - Predictive maintenance experience in CPG is a plus Required Skills: Machine learning, Feature engineering, Model evaluation, Python, pandas, scikit-learn, lifelines, AWS SageMaker, AWS Lambda, AWS Step Functions, AWS DynamoDB, AWS S3, Survival analysis, Weibull AFT, Cox PH, Competing risks, Remaining Useful Life (RUL), Predictive maintenance, Operational data processing, Sensor data analysis, ETL pipelines, Production ML deployment Benefits: Meaningful equity, Claude Code access, W2 option