Manraj Sidhu

Data Scientist @Redeployable

Built a collaborative filtering recommendation system using cosine similarity algorithms that processes client data across multiple features, generating personalised service recommendations while excluding already-used servicesDeveloped machine learning prioritisation models using weighted scoring algorithms with feature importance analysis, implementing risk-based scoring that identifies high-priority clients through protective and risk factor weighting based on logistic regression coefficientsImplemented medallion architecture data pipeline with bronze (raw data), silver (processed), and gold (analysis-ready) layers, ensuring data quality through automated validation, feature engineering, and comprehensive data quality frameworksCreated feature engineering pipelines that transform raw client data into analytical features including engagement scores, employment barriers, professional certifications, and skills assessments, enabling sophisticated similarity matching and priority scoringDesigned automated scoring algorithms that calculate priority scores using both demographic and non-demographic features, with categorical variable handling and numerical feature standardisation for fair comparisons across diverse client populations