Harnitha Suresh.

WOW! I am thrilled you're here! Thanks for reaching out!

Feel free to explore my projects created during my free time.

ML & AI

Thumbnail Analyser

EDA Descriptive Analyzer

Research Paper Analyser

Digit recognizer

ML Project

Thumbnail Analyser

A project aimed to explore Neuro-Marketing.

In midst of scrolling and the click of a video thumbnail, what makes us click it.

Well, the brain sees through the eyes ,but what grabs the attention of the eyes?. the colors, the contrast, the image placement,....,
in short - what stands out!
in other terms saliency.

This project analyses the saliency points of an image using opencv, saliency_map, numpy and pillow libraries.
This tool not only works for Youtube thumbnails but for any image.

Go check where your attention goes.

Images

EDA Descriptive Analyzer

A project aimed to expand scope of df.describe() function in pandas dataframe.

Understanding the data is the one of the most crucial and intial steps in making models.

Presenting a simple, select and analyse ,no code tool that can drop columns, find categorical features, provide complete descriptive statistical information for each column in the dataset.

Just upload your csv file and play around.

Libraries Used: streamlit, pandas

Images

Research Paper Analyser

A project aimed to explore LLMs and NLP techniques

Research paper is an integral part in every masters student, skimming and understanding long papers can be exhaustive.

With a goal in mind to make the prioritization of which papers to read from 10s of bookmarks ones, this tool is aimed to help achieve a good high level understanding of the uploaded paper.

Currently, it supports summarization using pretrained summarization models from transformers library.

Future implementation shall include fine-tuning the model with the paper's content and a chatbot to question and answer your queries.

Libraries used: transformers, PyPDF2, streamlit.

Model and Tokenizer used: BartTokenizer and BartForConditionalGeneration of facebook/bart-large-cnn model

Images

Digit recognizer

A project aimed to explore CNN classification

A fun tool that predict the digit you have drawn.

The prediction model is a classic CNN model trained on famous MNIST digit dataset with a accuracy score of 96%.

The project includes image preprocessing of the user drawn digit inorder to feed in the model for classification.

Give it a try, did the model guess your number right?

Libraries used: streamlit, numpy, pillow, TensorFlow.keras, tensorflow

Images

Predicting PCR and RFS in Breast Cancer

A university coursework project.

Images

Time Buddy

A daily meeting calender, created, accessed and modified using a NLP chatbot.

used nltk library for text-preprocessing and similarity-based intent matching.

Front-end in progress, coming soon... A university coursework project.

Images

PowerBI

Image

Harnitha Suresh

Masters Student at University of Nottingham

Masters in Data Science

Isn't it mind-blowing to learn about how the brain works—how we guess, analyse, decide, generate, observe, and learn. And if there's a field that actually tries to mimic it, what could be more exciting than signing up for it?

Hi, I'm Data Science enthusiast, a curious explorer of neuroscience, and I'm passionate about understanding how to code the brain's gifts through machine learning and artificial intelligence. Generative AI always excites me, but enjoy data analysis, classification and clustering too.

I started my software engineering journey with React and PowerBI, accumulating 2 years of industry experience. Now, I'm stepping into the world of AI through a Master's in Data Science. I'm proficient in PowerBI, SQL, and Python, and I've had hands-on experience with libraries such as Scikit-learn, NumPy, Pandas, OpenCV, TensorFlow.keras, Matplotlib, NLTK, Streamlit, Transformers, and Apache Spark. Currently, I'm exploring Langchain, Stable Diffusion, and model fine-tuning.

Contact me through LinkenIn

Copyright ©. All Rights Reserved. — Designed with love by Untree.co Distributed By ThemeWagon