Projects

You can check out some of my past work here! Feel free to contact me with ideas and suggestions :)

PS - Some of my projects are linked to repositories/work examples!

Implemented and compared several regression machine learning models for popularity prediction of songs.
Performed data mining, exploration, augmentation, and cleaning tasks to compile the dataset for analysis.
Evaluated the impact of features on predictions using Feature Importance and SHAP plots.

Comparative Analysis of SAC and PPO Algorithms on Walker2D Environment

Applied and compared Reinforcement Learning algorithms on the Walker2D Mujoco environment.
Implemented Soft-Actor-Critic and Proximal Policy Optimization algorithms in Python.

Designed and implemented a low-cost vision-based system to intelligently segregate recyclable waste.
Applied TinyML principles to execute a compressed neural network for object recognition using Python.
Interfaced a webcam module with a Raspberry Pi 3B to build the hardware pipeline.

Predicting survival of patients with heart failure using Machine Learning and Survival Analysis

Analyzed data by applying biostatistical and ML methods for feature ranking.
Extracted the most important risk factors to be used in predictions.
Compared the results obtained by applying various ML algorithms before and after feature ranking using R.

GPU Acceleration of Fractal Generation Algorithms

Improved the performance of the Escape Time Algorithm for fractal generation by parallelizing it.
Applied the Mariani/Silver algorithm to further improve the algorithm.
Implemented in Python leveraging PyCuda library for kernel design.

Single-Image Depth Perception in the Wild

Implemented an algorithm for single-image depth perception in the wild, using the “Depth in the Wild” dataset.
Incorporated an algorithm to estimate metric depth using annotations of relative path.

Smart Mirror

Designed and built a low-cost smart mirror using Raspberry Pi 4, Pi camera and Python Imaging and Deep Learning libraries.
Implemented real-time Skeleton Tracking with Augmented Reality for users to choose and view clothing options from their own wardrobe.

Ideated and pitched a transit app to track a user's carbon footprint based on their chosen mode of transportation.
Conducted and incorporated customer-based research into the design to provide the best user experience.
Performed market research and analysis to create a unique product to promote sustainability in New York City.