B.Tech Student at University of Pune | GSSoC 2025 | Aspiring Data Scientist | Passionate about AI, ML & Data Science

Aayush Kumbharkar  • Peerlist

Built and deployed an AI-enhanced web application that analyzes resumes against job descriptions to identify missing keywords and optimize for Applicant Tracking Systems (ATS). The tool leverages OpenAI’s LLMs and natural language processing to provide smart suggestions, helping candidates tailor their resumes for higher job-matching accuracy. 🛠 Key Features: Compare any resume with a job description to highlight missing and matched keywords. Generate AI-powered suggestions to improve keyword alignment. Easy-to-use interface deployed with Streamlit. Hosted on Streamlit Cloud for instant access and zero installation. 🔍 How it works: Users simply upload their resume (PDF/Text) and paste the job description. The app performs a semantic comparison, returning actionable insights for improving the resume's compatibility with ATS filters. 🌐 Live App: https://pro-ats.streamlit.app/

AI-Powered Resume Keyword Analyzer | ATS Opti

Analyzed PM2.5, NO₂, and CO₂ datasets from public APIs and open data sources to uncover pollution hotspots, seasonal variations, and long-term emission trends. Gathered & processed large-scale air quality datasets ensuring data consistency and accuracy. Conducted exploratory data analysis (EDA) to identify key environmental patterns. Built interactive Tableau dashboards enabling geographic emission benchmarking for policy research & environmental monitoring. Created data visualizations in Python to complement dashboard insights.

Air Quality & Emissions Data Analysis

Built an object detection system using the YOLOv8 model from Ultralytics, applied directly to images, video files, and real-world objects using a live webcam. What I did: Integrated YOLOv8 with OpenCV for real-time object detection without custom training. Detected objects from the COCO dataset with bounding boxes, class names, and confidence scores. Used YOLOv8n for speed and YOLOv8l for higher accuracy, depending on the use case. Added FPS monitoring and clean visual output using cvzone. Technologies used: Python, OpenCV, Ultralytics YOLOv8, cvzone This project gave me practical experience with real-time deep learning inference, computer vision pipelines, and video stream processing. GitHub Repository: https://github.com/aayushkumbharkar/YOLOv8-Object-Detection

End-to-End Object Detection System using YOLO

The project enabled us to develop skills in Arduino programming, sensor integration, and data handling. We gained a deeper understanding of how sensors work and how they can be utilized in different applications such as environmental monitoring, home automation, and industrial control systems.

Microproject on creating a dataset using an A

My NVIDIA Stock Price Prediction Bot Application on Streamlit Cloud Services. This application utilizes machine learning techniques to predict stock prices and provides insights into stock trends. Users can input stock ticker symbols and receive predictions, technical indicators, and even generative AI-based stock reports. 🔗 Check it out here: https://nvidia-stock-prediction-bot.streamlit.app/ Key Features: Real-Time Stock Data: Fetches historical stock data using the Yahoo Finance API. Technical Indicators: Calculates and visualizes key indicators like moving averages and Relative Strength Index (RSI). Stock Price Predictions: Utilizes a Linear Regression model for future stock price predictions. Generative AI Reports: Generates insightful reports on stock performance using pre-trained GPT models.