ML Engineer (CE50SF RM 4206)

Position: ML Engineer (CE50SF RM 4206)

Shift timing : 02 PM to 10 PM
Work Mode : Work From Office
Required Industry Experience : 5+ years of total development experience
Relevant Experience required : 2.5+ years of relevant Gen AI experience
Education Required: Bachelor’s / Masters / PhD: Bachelor’s degree in engineering

Must have skills:

• Python for ML
• Scikit-learn, PySpark
• Supervised Learning – Logistic Regression, Random Forest, XGBoost/LightGBM/CatBoost, calibration
• Unsupervised Learning & Anomaly Detection, Imbalanced Data Techniques – Class weighting, focal loss, threshold tuning, cost-sensitive learning
• Model Evaluation & Calibration, Statistics & Probability
• Drift Monitoring – Data drift, concept drift, PSI, KL divergence, model performance monitoring, retraining strategies
• ML Architecture Design – End-to-end pipelines
• Databricks – Notebooks, Delta Lake, Jobs, Workflows, Feature Store, MLflow integration, Unity Catalog,
  MLOps Fundamentals

Good to have skills:

• Experience in Azure
• Deep Learning – PyTorch / TensorFlow, sequence models (LSTM/Transformers) for fraud detection
• Git-based workflows (branching, pull requests, code reviews) and Agile/Scrum delivery

Any special or skills related notes

• Hands-on and accountable for delivering working, supportable solutions
• Clear communicator who can translate between business needs and technical implementation
• Quality-focused (testing, monitoring, documentation) with attention to reliability and maintainability
• Calm under pressure when responding to incidents and prioritizing production work
• Ownership & accountability across the full ML lifecycle
• Mentoring other team members / knowledge sharing

Role focus: Core ML modelling
Key responsibilities

• Design ML architecture (feature store, training pipelines, scoring)
• Build:
  o Supervised diversion-risk model
  o Unsupervised anomaly detection model
  o Model evaluation, calibration, and drift monitoring
• Define reusable feature engineering framework
• Design response schema:
  o Risk score
  o Explainability layer

Required skills

• Strong Python (scikit-learn, PySpark)
• Experience with anomaly detection techniques
• Experience with imbalanced datasets (fraud/risk domains preferred)
• Knowledge of MLOps (Databricks preferred)

************************************************************************************