Handwriting to Speech Recognition

Facilitated IAM database to train traditional ML models that takes words from a handwritten note as ‘bounding boxes’ & redirects through a voice assistant that reads out the text via Artificial Intelligence/Machine Learning frameworks helping the visually impaired population with an improved perform

Big Data to predict crime Insights

The problem of crime prediction in Chicago is critically important due to its potential to enhance public safety & resource allocation. The project utilizes a diverse dataset that includes historical crime data for Chicago containing information such as crime types, locations, timestamps.

Spark Data Processing & Streaming

Collected stream of log entries to analyze traffic, using data structures & algorithms to implement on domain extracting useful information, trends and patterns.

Data Ingestion Platform

Python based Open Source ETL tool for file crawling, document processing, content analysis, data ingestion to create a data lake to reduce processing time by 30%.

AI - ChatBot

A tailored medical chatbot, integrating SQL databases and NLP capabilities, to deliver precise and valuable information to patients; achieved a 70% reduction in repetitive inquiries & self-service interaction by 65%.

Income Prediction

Incorporated data mining techniques to predict income of certain groups of people to be above $50k/per year and presented insights to clients for improved costs and strategies.

Bayesian Regressions:Comprehensive analysis of the models utilized in Clumped Isotope Thermometry

The project aims to contribute to the scientific understanding of carbonate clumped isotope thermometry by investigating and evaluating various regression models, so we can get an idea of the best robust model under different conditions (e.g. error scenario, number of observations).

AI - Extended Sudoku Project

We use backtracking algorithm for solving and generating standard 9x9 Sudokus puzzles and add an extra functionality. First we do Backtracking and then we add a functionality where both diagonals need to contain all numbers from 1 to 9 without repetition.

SQL Data Exploration

A data exploration regarding a COVID-related dataset conducted through Microsoft SQL Server.

Bike Sales Dashboard

A dashboard made with Sheets that shows the amount of bikes purchased with respect to customer commute distance as well as various demographics.

Paldean Pokémon Stat Tier Generator

AI-Driven Histopathology Analysis for Early Cancer Detection

Develop an advanced artificial intelligence (AI) system that leverages deep learning techniques to detect and classify cancerous cells in histopathology images at an early stage to reduce diagnostic times and increase the accuracy of cancer diagnoses by automating the detection process.

Breast Cancer Detection using Deep Learning

This project focuses on detecting breast cancer from histopathology images using deep learning techniques. It utilizes the Fastai library to improve the predictive performance and provides a comprehensive set of Jupyter notebooks detailing the experiments.