Sagarika Kandula

About

Highly analytical Computer Science student specializing in Artificial Intelligence and Machine Learning, with hands-on experience in machine learning, data analytics, computer vision, and NLP. Builds end-to-end ML solutions and interactive dashboards using Python, SQL, and common data science tools.

Education

• B.Tech
  ACE Engineering College
  Computer Science and Engineering in Artificial Intelligence and Machine Learning · 2023 – 2027

Skills

Tools / apps / platforms

Projects

• Amazon Sales Analytics & Business Intelligence
  Python, SQL, Power BI, Pandas

  Analyzing Amazon sales data using Python and SQL to identify revenue trends, product performance, customer behavior, and profitability. Performing data cleaning, transformation, and exploratory data analysis (EDA) to prepare the dataset and uncover actionable business insights. Developing an interactive Power BI dashboard to track key performance indicators, sales trends, and profitability metrics. Applying customer segmentation and sales trend analysis to support data-driven business decision-making.

• Flight Delay Prediction using Machine Learning
  Python, Pandas, Scikit-learn, XGBoost

  Developed a dual ML pipeline to predict flight delay duration and delay probability using flight and weather data from 20,000 synthetic records. Performed EDA and feature engineering, incorporating weather conditions, peak-hour congestion, seasonal patterns, and prior flight delays. Trained and tuned XGBoost regression and classification models achieving 8.71 min MAE for delay duration and 0.789 ROC-AUC for delay classification. Applied SHAP explainability and probability calibration to identify key delay drivers including severe weather, inbound delays, and evening peak departures.

• Customer Churn Prediction using Machine Learning
  Python, Pandas, Scikit-learn, XGBoost, SQL

  Built an end-to-end churn prediction pipeline on 7,000 telecom customer records, including data cleaning, EDA, feature engineering, preprocessing, and model evaluation. Compared Logistic Regression, Random Forest, Gradient Boosting, and XGBoost using Accuracy, Precision, Recall, F1-score, and ROC-AUC. Tuned an XGBoost classifier using GridSearchCV and achieved 0.789 ROC-AUC, with optimized thresholding improving churn recall to 64.3%. Identified key churn drivers and developed actionable customer retention strategies.

• AI-Powered Exam Proctoring System
  Python, Flask, OpenCV, MediaPipe, SQLAlchemy, NumPy

  Developed a real-time AI-assisted proctoring system using MediaPipe and OpenCV for webcam-based face and behavior analysis. Implemented multi-face detection, head-pose estimation, LBPH face verification, and object detection for phones, laptops, and books with configurable detection thresholds. Built automated violation detection and evidence capture for suspicious behavior, missing/unrecognized faces, multiple people, and prohibited objects. Integrated Flask, SQLAlchemy, and SQLite to manage exams, sessions, warnings, violation logs, evidence, and admin monitoring.

Courses & certifications

• SQL and Database Management Systems with Project · EduSkills Academy · 2026
• SQL for Data Analysis · LinkedIn Learning · 2025
• Advanced Python · LinkedIn Learning · 2025