Senior Data Scientist – Manufacturing Intelligence, Machine Learning & AI

We are looking for a Senior Data Scientist to help build advanced analytics, machine learning, and AI solutions for manufacturing operations. 

This role will focus on using factory data to detect anomalies, improve quality, reduce downtime, optimize throughput, and support reusable data models that connect fragmented manufacturing systems into a common intelligence layer. 

The ideal candidate has strong applied machine learning skills, practical experience working with complex operational data, and the ability to partner with manufacturing, data engineering, platform, and software teams to move analytical solutions toward production. 

This is not a pure research role. We are looking for someone who can move from problem framing to data understanding, model development, validation, stakeholder alignment, and production support. The candidate should be able to learn unfamiliar domains quickly, challenge assumptions constructively, and push back when requirements, data quality, or model expectations are not realistic. 

Manufacturing experience is strongly preferred, but we are also open to candidates from adjacent industrial, operations, quality, aerospace, semiconductor, supply chain, or equipment-heavy environments who can learn the manufacturing domain quickly. 

Summary of Data Science Work in a Manufacturing Environment 

A Data Scientist in manufacturing works at the intersection of factory operations, engineering, quality, maintenance, data platforms, and machine learning. 

The work is not only about building models. It includes understanding how the plant operates, identifying where data is generated, defining what “normal” and “abnormal” look like, creating reliable features from machine and process signals, validating model outputs against real-world outcomes, and delivering insights that plant teams can act on. 

Typical manufacturing data science work includes detecting process drift, identifying abnormal machine behavior, predicting quality issues, improving equipment health visibility, supporting root cause analysis, and helping teams move from reactive firefighting to proactive detection, triage, and prevention. 

Success requires technical depth, manufacturing curiosity, practical judgment, and the ability to build solutions that work with messy, incomplete, noisy, and high-frequency industrial data. 

Key Responsibilities 

Applied Machine Learning & Analytics 

Develop machine learning and statistical models to support manufacturing use cases such as anomaly detection, quality prediction, equipment health, process monitoring, throughput improvement, and decision support. 

Apply supervised, unsupervised, and semi-supervised learning methods, including classification, regression, clustering, anomaly detection, time-series analysis, statistical process control, and model explainability. 

Build anomaly detection solutions using methods such as control limits, isolation forests, clustering, Mahalanobis distance, autoencoders, time-series models, and supervised classification where labeled defects are available. 

Develop models for manufacturing use cases such as stamping split detection, weld quality, paint defects, assembly issues, predictive maintenance, bottleneck detection, process optimization, and quality prediction. 

Evaluate model performance using appropriate metrics, ground truth definitions, validation strategies, false positive and false negative analysis, and business impact measures. 

Identify when data is insufficient, labels are unreliable, ground truth is weak, or a machine learning approach is not appropriate, and communicate those limitations clearly. 

Manufacturing Data & Feature Engineering 

Analyze real-time and historical factory data from sources such as PLCs, sensors, machines, MES, SCADA, historians, quality systems, maintenance systems, production logs, and enterprise platforms. 

Create features from manufacturing signals such as cycle time, pressure, temperature, torque, vibration, current, force, cushion pressure, line speed, JPH, FTT, FRC, scrap, rework, downtime, and fault codes. 

Work with noisy, incomplete, high-frequency, or fragmented industrial data to create reliable analytical datasets. 

Build features that reflect manufacturing context, including asset hierarchy, station behavior, part flow, process sequence, shift patterns, tool usage, maintenance history, supplier variation, and quality outcomes. 

Partner with plant teams and domain experts to understand process behavior, validate assumptions, and determine whether model outputs reflect real operating conditions. 

Cloud, Data Pipelines & MLOps 

Use cloud data platforms, preferably GCP, to support scalable analytics and machine learning workflows. 

Develop and partner with Data Engineering to build data pipelines that ingest, transform, and prepare manufacturing data for analysis, modeling, monitoring, and reporting. 

Work with tools such as BigQuery, Cloud Storage, Pub/Sub, Dataflow, Vertex AI, Cloud Run, Cloud Functions, Looker, or similar cloud services. 

Support real-time and near-real-time analytics use cases by working with streaming data from MQTT, Kafka, Pub/Sub, Dataflow, or similar event-driven architectures. 

Partner with platform and software engineering teams to move models and analytical workflows from prototype to production-ready solutions. 

Follow MLOps practices such as experiment tracking, model versioning, model deployment, model monitoring, drift detection, retraining workflows, and production documentation. 

Monitor model performance after deployment, including false positives, false negatives, data drift, model drift, latency, uptime, pipeline failures, and changing manufacturing conditions. 

Semantic Modeling & Manufacturing Context 

Contribute to manufacturing ontology, semantic modeling, and data relationship work by helping define consistent entities, features, metrics, and relationships across manufacturing systems. 

Support mapping of fragmented factory data sources into connected data models that can support analytics, reporting, AI agents, decision support, and reusable data products. 

Help create consistent definitions for manufacturing metrics such as JPH, OEE, FTT, FRC, downtime, scrap, rework, cycle time, takt time, throughput loss, and bottleneck impact. 

Partner with data architects, ontology leads, data engineers, and domain experts to validate that semantic models reflect how the factory actually operates. 

The candidate does not need to be an ontology or knowledge graph expert on day one, but should be willing and able to learn semantic modeling concepts and apply them to manufacturing analytics. 

Productization, Communication & Delivery 

Build dashboards, alerts, APIs, and explainability views that help plant teams understand what changed, why it matters, and what action to take. 

Develop lightweight internal applications or decision-support tools using frameworks such as Streamlit, Dash, FastAPI, Looker, Power BI, Grafana, or similar tools. 

Collaborate with data engineers, platform engineers, software engineers, manufacturing engineers, quality teams, and plant stakeholders to move data science prototypes into production-ready workflows. 

Follow software engineering best practices, including version control, modular code, code reviews, testing, logging, documentation, reusable packages, and reproducible environments. 

Document model logic, assumptions, input features, thresholds, limitations, operational dependencies, and recommended actions for business and plant-floor users. 

Distinguish between exploratory research, prototype development, and production-ready delivery. 

Prioritize high-impact use cases and help define practical delivery paths that can scale across plants, lines, stations, and manufacturing domains. 

Required Qualifications 

Bachelor’s or Master’s degree in Data Science, Computer Science, Statistics, Industrial Engineering, Mechanical Engineering, Manufacturing Engineering, Operations Research, Applied Mathematics, or a related technical field. 

5+ years of experience applying data science, machine learning, statistical modeling, optimization, or advanced analytics in a professional environment. 

Strong Python skills using libraries such as pandas, NumPy, scikit-learn, SciPy, XGBoost, PyTorch, TensorFlow, statsmodels, or similar tools. 

Strong SQL skills and experience working with large, complex datasets. 

Experience with supervised and unsupervised machine learning methods, including classification, regression, clustering, anomaly detection, time-series analysis, forecasting, or process optimization. 

Experience building features from machine, sensor, process, quality, maintenance, production, or operational datasets. 

Experience working with cloud-based data and analytics platforms such as GCP, AWS, Azure, or similar environments. 

Experience working with data engineering, software engineering, or platform teams to move analytical solutions toward production. 

Understanding of MLOps concepts such as experiment tracking, model deployment, model monitoring, CI/CD, version control, testing, model registry, and retraining. 

Ability to work with noisy, incomplete, high-frequency, or fragmented operational data. 

Ability to communicate technical findings clearly to plant teams, engineers, leaders, and non-technical stakeholders. 

Ability to operate in ambiguous environments where requirements, data quality, and success criteria may need to be clarified. 

Professional confidence to challenge assumptions, push back constructively, and influence stakeholders with evidence. 

Demonstrated ability to learn new technical and business domains quickly. 

Preferred Qualifications 

Experience applying data science or machine learning in manufacturing, industrial, automotive, aerospace, semiconductor, supply chain, quality, maintenance, or operations environments. 

Experience with automotive manufacturing, stamping, body shop, paint shop, final assembly, battery manufacturing, or powertrain operations. 

Understanding of manufacturing KPIs such as throughput, cycle time, downtime, OEE, JPH, FTT, FRC, scrap, rework, takt time, bottlenecks, quality escapes, and safety events. 

Basic understanding of manufacturing systems such as MES, SCADA, PLCs, historians, CMMS, QMS, ERP, or industrial IoT platforms. 

Hands-on experience with GCP services such as BigQuery, Cloud Storage, Pub/Sub, Dataflow, Vertex AI, Cloud Functions, Cloud Run, Looker, Cloud Build, Artifact Registry, or Cloud Monitoring. 

Experience with real-time or near-real-time anomaly detection on streaming manufacturing data. 

Experience with MQTT, Kafka, GCP Pub/Sub, AWS IoT, Azure IoT, Spark, Dataflow, Flink, or similar streaming and industrial IoT tools. 

Experience creating or using ontologies, semantic layers, knowledge graphs, data catalogs, data fabric platforms, or entity-relationship models. 

Familiarity with graph databases or semantic technologies such as RDF, OWL, SPARQL, Neo4j, Stardog, GraphDB, or similar tools. 

Familiarity with manufacturing data standards, asset hierarchies, ISA-95, OPC UA, industrial data models, data mesh, semantic layers, or digital twin concepts. 

Experience building production ML systems with automated training pipelines, validation gates, deployment workflows, monitoring, rollback strategies, or retraining triggers. 

Experience with containers and deployment platforms such as Docker, Kubernetes, GKE, Cloud Run, or managed inference endpoints. 

Experience building explainable AI outputs for operational users, including root-cause indicators, contributing factors, confidence scores, thresholds, and recommended actions. 

Experience with visualization tools such as Power BI, Tableau, Looker, Grafana, Streamlit, Dash, or custom web applications. 

Success Measures 

Success in this role will be measured by the ability to: 

Reduce quality defects, scrap, rework, downtime, and throughput loss through better detection and prediction. 

Improve visibility into manufacturing performance, process drift, equipment behavior, and root causes. 

Build reusable models, features, and data products that can scale across plants, lines, stations, and manufacturing domains. 

Support trusted manufacturing data definitions, ontology structures, and semantic relationships. 

Partner effectively with data engineering, software engineering, platform, and manufacturing teams to deploy reliable analytics. 

Improve model reliability through monitoring, drift detection, retraining, and production documentation. 

Deliver analytics that plant teams can understand and act on quickly. 

Bridge the gap between manufacturing domain experts, data science, data engineering, software engineering, and platform teams. 

Ideal Candidate Profile 

The ideal candidate is a strong applied data scientist who is comfortable with both the technical and operational realities of manufacturing. 

They are strong across supervised and unsupervised machine learning, practical and delivery-oriented, comfortable with messy real-world data, and able to understand manufacturing systems and plant-floor context. 

They can translate vague operational problems into structured analytical approaches, explain model results and limitations clearly, and push back constructively when requirements, data quality, or success criteria are not realistic. 

They do not need to be an expert in every area on day one. The strongest candidate will bring deep applied machine learning capability, sound technical judgment, professional confidence, and the ability to learn manufacturing, cloud, MLOps, and semantic modeling concepts as needed.