Posted 26 July, 2026
Specialist-Machine learning
NR Consulting
Pune,Maharashtra
Full Time
Reference: 365_463738_26-21907
Title: Specialist—Machine learning
Location: Pune
Exp: 5-9 Years
Job Description:
Key Responsibilities
• Design and train the Gate Readiness Prediction model: classify/predict likelihood of gate success across all 6 gates (Pre-KO → SOVP) using historical SNPD gate cycle times, deliverable completion rates, and predecessor adherence.
• Build the Predictive Risk Heat-map: compound urgency risk scoring across project, part, and supplier dimensions using structured SNPD, SAP cost, and PLM data.
• Develop the Supplier Risk Scoring model (Phase 2): supplier profiling using PPAP outcomes, AQP scores, vendor quality history from SAP, and onboarding benchmarks.
• Build the Lessons-Learned Recommender (Phase 2): retrieval + ranking model over historical project data, gate outcomes, and documented lessons in the SNPD knowledge base.
• Engineer training feature pipelines from the three-tier data architecture: Azure SQL (live), ADLS Gen2 Delta Lake (historical/ML features), and Azure AI Search (vector).
• Implement SHAP-based explainability outputs for all predictive models — mandatory for gate reviewer trust and IATF 16949 audit traceability.
• Set up and maintain MLflow experiment tracking, model registry, and deployment versioning on Azure ML.
• Integrate model inference endpoints into SNPD AI microservices; ensure low-latency serving (P95 < 2s for gate readiness queries).
• Implement model monitoring: drift detection, performance degradation alerts, and automated retraining triggers via Azure Data Factory pipelines.
• Conduct rigorous model evaluation: precision/recall/F1 for classification tasks, MAE/RMSE for regression; fairness checks across project types (M2/M4/M6) and portfolio buckets.
• Collaborate with the Data Engineer on feature store design, training data quality, and CDC-based incremental refresh from SAP/PLM sources.
• Document model cards: training data lineage, performance benchmarks, known limitations, and re-training cadence for each deployed model.
TECHNICAL SKILLS REQUIRED
GOOD TO HAVE
• Experience with automotive or manufacturing NPD data (DFMEA, DVP, PPAP, BOM structures).
• Familiarity with SAP S/4 HANA data models (Material Master, BOM, Vendor, Cost Center).
• Knowledge of Teamcenter PLM part lifecycle and ECN data structures.
• Deep learning with PyTorch / TensorFlow for NLP or document understanding tasks.
• Databricks / Apache Spark for large-scale feature computation.
• IATF 16949 quality management system awareness.
Location: Pune
Exp: 5-9 Years
Job Description:
Key Responsibilities
• Design and train the Gate Readiness Prediction model: classify/predict likelihood of gate success across all 6 gates (Pre-KO → SOVP) using historical SNPD gate cycle times, deliverable completion rates, and predecessor adherence.
• Build the Predictive Risk Heat-map: compound urgency risk scoring across project, part, and supplier dimensions using structured SNPD, SAP cost, and PLM data.
• Develop the Supplier Risk Scoring model (Phase 2): supplier profiling using PPAP outcomes, AQP scores, vendor quality history from SAP, and onboarding benchmarks.
• Build the Lessons-Learned Recommender (Phase 2): retrieval + ranking model over historical project data, gate outcomes, and documented lessons in the SNPD knowledge base.
• Engineer training feature pipelines from the three-tier data architecture: Azure SQL (live), ADLS Gen2 Delta Lake (historical/ML features), and Azure AI Search (vector).
• Implement SHAP-based explainability outputs for all predictive models — mandatory for gate reviewer trust and IATF 16949 audit traceability.
• Set up and maintain MLflow experiment tracking, model registry, and deployment versioning on Azure ML.
• Integrate model inference endpoints into SNPD AI microservices; ensure low-latency serving (P95 < 2s for gate readiness queries).
• Implement model monitoring: drift detection, performance degradation alerts, and automated retraining triggers via Azure Data Factory pipelines.
• Conduct rigorous model evaluation: precision/recall/F1 for classification tasks, MAE/RMSE for regression; fairness checks across project types (M2/M4/M6) and portfolio buckets.
• Collaborate with the Data Engineer on feature store design, training data quality, and CDC-based incremental refresh from SAP/PLM sources.
• Document model cards: training data lineage, performance benchmarks, known limitations, and re-training cadence for each deployed model.
TECHNICAL SKILLS REQUIRED
| • Python 3.10+ — scikit-learn, XGBoost, LightGBM | • Azure Data Factory (pipeline orchestration for ML) |
| • Time-series forecasting (project timeline prediction) | • ADLS Gen2 / Delta Lake (feature store reads) |
| • Azure Machine Learning (training, registry, endpoints) | • SQL Server / T-SQL (SNPD unified data model) |
| • MLflow — experiment tracking & model versioning | • Azure AI Search (embedding-based retrieval) |
| • SHAP / LIME — model explainability | • Docker / Azure AKS for model serving |
| • Feature engineering from SQL + Delta Lake | • pandas, NumPy, Plotly (analysis & visualization) |
GOOD TO HAVE
• Experience with automotive or manufacturing NPD data (DFMEA, DVP, PPAP, BOM structures).
• Familiarity with SAP S/4 HANA data models (Material Master, BOM, Vendor, Cost Center).
• Knowledge of Teamcenter PLM part lifecycle and ECN data structures.
• Deep learning with PyTorch / TensorFlow for NLP or document understanding tasks.
• Databricks / Apache Spark for large-scale feature computation.
• IATF 16949 quality management system awareness.
