Engineering student focused on machine learning, data science, and applied research. Building end-to-end ML systems, exp

Janani Ravi  • Peerlist

I recently completed a project where I built a machine learning pipeline to predict delivery times for e-commerce orders. The goal was to improve logistics efficiency and enhance customer satisfaction by accurately estimating delivery durations based on factors like product category, distance, traffic, weather, and agent performance. What I did: - Cleaned and preprocessed a detailed delivery dataset - Performed Exploratory Data Analysis (EDA) to uncover insights on delivery trends, agent performance, and external conditions - Engineered features like geospatial distances and time-based variables - Developed regression models: Linear Regression, Random Forest, Gradient Boosting - Tracked and compared models using MLflow - Built a Streamlit app for real-time delivery time predictions Business Impact: - Optimize delivery schedules - Adjust delivery estimates dynamically for traffic & weather - Evaluate delivery agent performance - Enhance customer satisfaction Tech Stack: Python | scikit-learn | LightGBM | Streamlit This project strengthened my skills in data cleaning, EDA, feature engineering, regression modeling, and end-to-end deployment. 💡 Excited to apply these skills to real-world logistics and e-commerce challenges!

📦 Amazon Delivery Time Prediction

Built an AI powered web application that predicts brain tumor types from MRI scans using deep learning. The model classifies images into four categories: -Glioma -Meningioma -Pituitary Tumor -No Tumor The system uses a combination of custom CNN models and transfer learning with InceptionV3, achieving 87.4% accuracy on the test set. The project leverages TensorFlow, Keras, and Scikit-learn for model training, evaluation, and preprocessing, including image augmentation and normalization. To make it interactive and user-friendly, the solution is deployed via Streamlit, allowing users to upload MRI images and instantly view predictions along with a confidence score and a class-wise probability bar chart. This tool supports medical professionals by offering an AI-assisted second opinion, potentially reducing diagnostic errors and enhancing early tumor detection. -Check out the app here :  https://brain-tumor-mri-image-classification-caw62ryadfwxmrhbutpuos.streamlit.app/

🧠 Brain Tumor Prediction Web App

I'm excited to share a project I recently completed where I built a machine learning-powered web app that predicts greenhouse gas emissions based on U.S. supply chain and industry data. 🧠 What it does: -Predicts GHG emissions using a trained ML model -Built with Python, scikit-learn, and Streamlit -Uses real-world data from U.S. industry emission factors -Deployed live using Streamlit Community Cloud This was a great hands-on experience working with real datasets, building an ML pipeline, and deploying a fully interactive web app! 🚀 👉 Check out the app: https://ghg-prediction-edunet-shell-internship-project.streamlit.app/