Mahesh Ponnam

Data Scientist

[email protected] +1 2679921203

Hello, I am

Mahesh Ponnam.

1

Machine Learning
Project

1

GenAI
Project

1

Deep Learning
Project

About Me

Hi there! I'm Mahesh Ponnam, a Data Scientist with 3 years of hands-on experience in Data Science and Analytics. I enjoy turning complex data into clear, actionable insights to solve real-world challenges, especially in healthcare and service.

Tools and Technologies:
➡️ I work with Python to clean data, build predictive models, and create machine learning solutions.
➡️ I'm comfortable using SQL for handling, querying, and managing large datasets.
➡️ I also use tools like Jupyter Notebooks, Pandas, Matplotlib, Scikit-learn, PyTorch, and Tableau to explore data, build models, and visualize results in a meaningful way.

Areas of Expertise:
➡️ Building machine learning models for things like healthcare premium prediction and credit risk scoring in finance.
➡️ Applying Natural Language Processing (NLP) and Generative AI to pull insights from text data and create smart, automated systems.
➡️ Automating data pipelines and ETL processes to make data workflows faster and more efficient.

Key Skills

Python

Machine Learning

Natural Language Processing (NLP)

Generative AI

SQL

Pandas

My Projects

Healthcare Premium Prediction Model(ML Project)

Domain/Function: Healthcare / Insurance

E-Commerce Chatbot with Semantic Routing

Domain/Function: E-Commerce | Conversational AI

Car Damage Classification using Deep Learning (ResNet50 & CNN

Domain/Function: Deep Learning

My Experience

Awards & Certificate

Math and Statistics For AI, Data Science

Python: Beginner to Advanced For Data Professionals

SQL for Data Science

Master Machine Learning for Data Science & AI: Beginner to Advanced

Deep Learning: Beginner to Advanced

Natural Language Processing

Let's Connect

Feel free to get in touch with me. I am always open to discussing new projects, creative ideas or opportunities to be part of your visions.

Mail Me

[email protected]

https://www.linkedin.com/in/maheshponnam3629?lipi=urn%3Ali%3Apage%3Ad_flagship3_profile_view_base_contact_details%3B%2Fwp3ploITx%2B5KNdtqPD0vA%3D%3D

Call Me

+1 2679921203

GitHub

https://github.com/MaheshPonnam6842

Machine Learning Project

Healthcare Premium Prediction Model(ML Project)

Domain/Function: Healthcare / Insurance

Insurance Amount Predictor: Machine learning model that calculates personalized insurance coverage amounts based on health profiles (age, BMI, habits). Built with Python to assist insurers and users in determining optimal coverage.

View Project
on GitHUB

View Project
on

Frontend for my insurance predictor: input health details, get instant premium estimates.

GitHub repo for healthcare insurance predictor ML model with Streamlit web interface. Includes Python code, helper files, and Apache license.

Project Details

Situation:
Insurance companies needed a reliable way to predict premiums based on customer risk factors.

Task:
Build a model to estimate healthcare insurance premiums using demographic and health data.

Action:
Cleaned and engineered features from healthcare data, then developed and fine-tuned ML models using Python, Scikit-learn with cross-validation, for robustness.

Result:
Delivered a model that improved premium prediction accuracy, enabling more personalized and competitive insurance pricing.

GenAI Project

E-Commerce Chatbot with Semantic Routing

Domain/Function: E-Commerce | Conversational AI

An AI-powered chatbot for e-commerce, deployed on Streamlit Cloud. It classifies queries into FAQs, product searches, and small talk using semantic routing. Tech stack: Python, Groq LLM, ChromaDB, SQLite, and Streamlit.

View Project
on GitHUB

View Project
on

Streamlit chatbot UI showing a product query for Puma shoes with the top 5 discounted results fetched from the SQLite database.

Code implementation in PyCharm demonstrating SQL chain logic, Groq API integration, and FAQ ingestion with ChromaDB

Project Details

This project is an AI-powered e-commerce chatbot deployed on Streamlit Cloud. It uses semantic routing to intelligently classify user queries into FAQs (policies, payments, shipping), SQL-based product searches (brand, price range, discounts), and small talk. The backend integrates ChromaDB for FAQ retrieval, SQLite for product data, and Groq LLM for natural language understanding and response generation. Built with Python, Streamlit, and HuggingFace sentence transformers, the chatbot provides a seamless and interactive experience for online shoppers. This project demonstrates skills in GenAI application design, NLP, vector databases, and cloud deployment.

Deep Learning Project

Car Damage Classification using Deep Learning (ResNet50 & CNN

Domain/Function: Deep Learning

A deep learning–powered application that classifies vehicle damage into Five categories (front/rear, breakage/crushed/normal). Built with CNN and ResNet50, deployed using Streamlit for real-time predictions from uploaded car images.

View Project
on GitHUB

View Project
on

Other Tools

scikit-learn

Pytorch

Matplotlib / Seaborn

Streamlit

GitHub repository showcasing project code, README, and implementation files for car damage predictor

Streamlit app interface where users upload a car image and receive instant damage classification.

Project Details

This project focuses on classifying vehicle damage images using Convolutional Neural Networks (CNN) and transfer learning with ResNet50. The dataset consists of 2300 car images categorized into six classes: Front_Breakage, Front_Crushed, Front_Normal, Rear_Breakage, Rear_Crushed, and Rear_Normal.

To improve model performance, extensive preprocessing was applied including data augmentation (random flips, color jitter, affine transforms) and normalization using ImageNet statistics. Three models were tested: a custom CNN, CNN with regularization (BatchNorm + Dropout), and ResNet50 fine-tuned on the last layer.

ResNet50 achieved the best results, showing strong performance across precision, recall, and F1-score. The final model was deployed as a Streamlit web app where users can upload car images to receive instant predictions.

This project highlights my skills in deep learning, model optimization, computer vision, and deployment of AI solutions with a clean, interactive UI.