LIFT

A 'Partnership Performance Analytics' Saas platform for program management and growth in the marketing industry.

I helped build LIFT from scratch, covering data science, user experience, user journey, UI, and product management; leading a team of 5 devs while also managing data engineering and analytics to create proprietary dashboards.

Shuffle

A cybersecurity SOAR platform for security teams to use and share security workflows to everyone's benefit.

I started working on Shuffle with web platform feature updates and transitioned into machine learning, to eventually leading the AI development efforts on the platform.

Boost Histogram

A python binding for Boost::Histogram and one of the fastest libraries for histogramming.

I worked on improving boost histogram library by adding a new accumulator, histogram comparison functionality, and several features that lead to new version release.

HANDWRITTEN EQUATION SOLVER

Python, TensorFlow, Keras, cv2, Numpy, Pandas, Matplotlib, OS, IPython, Google Colaboratory

I worked on a deep learning model for detecting, classifying and solving handwritten equations based on tensorflow and keras (modelling), numpy and pandas (statistical data handling), and cv2 (computer vision) among others to make the project functional.

BITCOIN MEMPOOL OPTIMIZATION

Python, Pandas, Numpy, Google Colaboratory

I worked on Bitcoin mempool's transaction Optimization where the net output is maximized, as a part of Summer of Bitcoin 2021 challenge; where I approached the problem through data science and statistics.

DECISION TREE ML CLASSIFIER

Python, SciKit Learn, Pandas, Matplotlib, Seaborn, Google Colaboratory

I created a decision tree based machine learning algorithm to test its classification effectiveness using Python and its libraries on a public dataset; reaching an accuracy of 96.67% on Google-Colab environment.

TIME SERIES PREDICTION

Python, SciKit Learn, Pandas, Matplotlib, Seaborn, Google Colaboratory

I worked on creating a stock market (time series) prediction model with input features that include adjusted closing price, adjusted volume, high low percentage and percentage change to predict future price of stock (with help of regression model created using Scikit Learn library) with 99+ % accuracy.

INSTAGRAM POST REACH PREDICTION

Python, SciKit Learn, Pandas, Matplotlib, Seaborn, Google Colaboratory

I worked on creating an Instagram post reach prediction ML model (based on multivariate regression) with input features that include followers and time/likes to predict likes/time of a particular post (created using Scikit Learn library).

FLARE'21 WEBSITE

HTML, CSS, JavaScript, JQuery, Boostrap, AOS Library

As the technical head of SnC, I worked on creating the website for Flare 2021 (one of top 3 annual events held at PDEU) from scratch using HTML-CSS-JS alongwith AOS JS library.