In God we trust
All others must bring data
![](images/portfolio pic1.png)
Welcome to my portfolio! My name is Harshit and I am a data scientist and software developer. This portfolio showcases some of my work and achievements, and serves as a testament to my skills, knowledge, and experience. Within these pages, you will find examples of my projects, as well as a summary of my educatioan and work experience. I am excited to share my portfolio with you and hope that it gives you a sense of my capabilities and passion for my field. Thank you for visiting!
The only true wisdom is in knowing you know nothing.
Utilizing a dataset of image-caption pairs from the Bloom Library, the model helps businesses better support their digital products and services in low resource NLP languages by augmenting the curation of alt text content, image tags, and potentially harmful user uploads.
The objective of this project is to increase user experience by making the search experience less noisy. This project utilizes image recognition and NLP technique to find best ensembled results.
Project aims to assist a company planning to launch a gaming app, to give recommendations on price of the app, the availability of in-app ads in the app, and in-app purchases in the app. This analysis is done using exploratory data analysis of the attributes, the correlation of scraped attributes with the number of downloads and hypothesis formulation.
This project utilizes crypto currency data, where archived data is procured from the CoinMarketCap website, and live streaming data is procured from Coinbase API. The goal for the project is to provide a live prediction of market close volume sold across cryptos and closing exchange rate to generate insights that lead to quick profits in the Crypto Currency market.
This project aims to find 10-K reports efficacy by measuring how quantitaive a document is complied by a company. This involves scraping all the 10-K documents since 1993 and extracting section 7 and 8 from each document using regular expression. The extraction of $ amounts is implemented by utilizing various NLTK methods and regular expressions.
A dynamic web app built on Django Framework that implements a efficient O(n^2) algorithm to effectively create combinations of products for optimal packing.
A comprehensive product that considers various factors such as employee's salary, time taken, market demand, raw materials etc. to estimate cost of a product. This product, coupled with sales forecasting, is an excellent tool for pricing a product.
An ERP system at heart, this product became an integral part of a parent product I was working on. A pure Angular project, this product tends to capture different process of product building with 78 steps and 9 different stakeholders thus streamlining product development.