Riddhi Mundle

Technology Job Simulation – Virtual Internship

DELOITTE | FORAGE · JUNE 2025

Practical consulting-oriented experience applying data engineering and analytics skills.

• •Analyzed telemetry data from factory machines using Python
• •Built a validated data pipeline ensuring consistent machine health data
• •Designed a real-time dashboard for 36 machines across 4 factories
• •Created a consulting-style proposal covering UI/UX layout, authentication workflow, implementation roadmap, and support plan
• •Worked on requirement analysis, backend logic, and client-ready documentation

ML Developer

CODENCE STUDIO · AUGUST 2025

Developed intelligent matching algorithms and profile suggestion system for a social dating platform using machine learning and pattern recognition.

• •Built matching algorithms using Scikit-Learn to analyze user preferences, behavior patterns, and compatibility metrics
• •Developed profile suggestion system that recommends potential matches based on ML models analyzing user interactions, preferences, and profile data
• •Implemented pattern recognition models to identify user preferences, interests, and behavioral patterns from profile data and interaction history
• •Created 'How We Work' feature explaining the matching algorithm's logic and ML-based recommendation system to users
• •Designed profile suggestion engine using similarity-based matching and content-based filtering techniques to improve match quality
• •Optimized matching algorithms for better performance, processing user data efficiently to provide profile suggestions
• •Worked on feature engineering to extract meaningful patterns from user profiles, preferences, and interaction data for better matching accuracy

Anomaly Detection System

Built a robust anomaly detection system using Scikit-Learn to identify outliers in time-series data. Implemented Isolation Forest and Local Outlier Factor algorithms to detect unusual patterns and anomalies in real-time sensor data. Performed comprehensive data preprocessing including handling missing values, normalization, and feature scaling. Evaluated model performance using precision, recall, F1-score, and confusion matrices. The system successfully identified anomalies with high accuracy, enabling proactive maintenance and reducing false positives through careful threshold tuning and feature engineering.

Classification Model Pipeline

Developed an end-to-end classification pipeline from data preprocessing to model deployment. Implemented feature engineering techniques including one-hot encoding, feature scaling, and handling imbalanced datasets. Built and compared multiple classification algorithms (Random Forest, SVM, Logistic Regression) using cross-validation. Created a Flask REST API for predictions with input validation and error handling. The pipeline includes model evaluation using precision, recall, F1-score, ROC-AUC curves, and confusion matrices. Deployed the model with basic version control and monitoring.

Data Visualization Dashboard

Created comprehensive interactive dashboards for exploratory data analysis combining Power BI and Python visualization libraries. Developed custom visualizations using Matplotlib and Seaborn to analyze data distributions, correlations, and trends. Built interactive Power BI dashboards with drill-down capabilities, filters, and dynamic charts for stakeholder presentations. Performed statistical analysis and created heatmaps, distribution plots, and time-series visualizations to uncover insights. The dashboards enabled data-driven decision making by providing real-time insights into key metrics and patterns.

ML Model Deployment Platform

Built a machine learning model deployment platform using Flask for serving ML models via REST API. Created endpoints for model inference, health checks, and basic performance tracking. Integrated MLflow for experiment tracking and model versioning. Implemented basic model deployment workflow with logging and error handling. Deployed the platform on AWS with simple containerization. The platform enables model versioning and basic monitoring for ML model serving.

Pattern Recognition System

Developed pattern recognition models for matching algorithms using machine learning approaches. Implemented models using Scikit-Learn and TensorFlow for pattern identification and feature extraction. Built models combining traditional ML algorithms with basic neural networks for improved accuracy. Performed feature engineering to extract meaningful patterns from raw data. Optimized model performance through hyperparameter tuning and cross-validation. The system identifies patterns and relationships in data for matching and classification tasks. Implemented basic inference capabilities for practical use.