Role Overview:
1. Data ingestion and quality:
Perform basic data cleaning, validation, time alignment, and documentation based on sensor data and event logs of the machine.
2. Feature exploration
Engineer simple health indicators and explore correlations between indicators and outcomes like aborted builds, rework, and alarms
3. Forecasting prototypes
Under guidance, prototype lightweight forecasting/baseline methods (e.g., moving averages, EWMA, AR baseline, simple classification thresholds)
Compare methods using clear metrics (e.g., precision/recall for early warning, lead time, false-alarm rate)
4. Visualization and monitoring
Build simple dashboards showing trailing indicators, predicted risk bands, and recent anomalies
Create concise reports that explain findings to technical and non-technical audiences
5. Experiment design
Help structure offline back tests and small A/B-style evaluations to assess alert usefulness
Document assumptions, data gaps, and improvement ideas
6. Collaboration and knowledge capture
Work with engineers and maintenance teams to understand failure modes and thresholds
Standardize templates for data dictionaries, feature lists, and evaluation summaries
Ideal Candidate:
Should be pursuing the course.
Required Qualifications
Bachelor’s student in Engineering, Data Science, Computer Science, Applied Math, or related field
Comfortable with basic statistics and time series concepts (trends, seasonality, moving averages)
Proficient with Excel or Google Sheets; exposure to a programming language (e.g., Python) from coursework or self-learning
Strong communication, organization, and teamwork skills
Interest in predictive maintenance, reliability, or analytics for manufacturing
Desired Qualifications (Nice to Have)
Basic Python data stack exposure (pandas, matplotlib/seaborn)
Intro knowledge of anomaly detection or forecasting concepts (e.g., z-scores, EWMA, AR/ARIMA at a high level)
Familiarity with additive manufacturing data types (sensor logs, alarms, maintenance records)
Experience with simple dashboards
Understand how forecasting and anomaly detection can improve uptime, quality, and maintenance planning
Gain hands-on experience with time series preprocessing, feature engineering, and baseline models
Learn to evaluate alert quality and communicate tradeoffs (lead time vs. false alarms)
Build practical dashboards and reports for stakeholders
Exposure to additive manufacturing process