I’m a Grad student at NYU Tandon pursuing a Masters’s degree in Computer Engineering with a focus on Deep Learning and High Performance Machine Learning. I’m passionate and driven about software engineering, software architecture, and overall software product management.

Personal and Academic Projects ⚙️
  1. Benchmarking and Training image classification model Resnet-18 on varying compute configurations
    Developed a modified Resnet-18 using PyTorch without Batch Normalization layers, and trained it on the CIFAR10 dataset. The training was done using different machines on the NYU Greene super compute with different numbers of CPU and GPU cores, and also fine-tuning hyperparameters like the number of Dataloader workers and optimizers.
  2. Multi-Lingual Review Classification and Inference using Transformers
    Developed and deployed a text classification model using Hugging Face Transformers, using a multi-lingual text dataset containing real-life product reviews from Amazon.com. The model was deployed on the Hugging face model hub and later utilized for inference
  3. Secure Chat Application
    Developed an Andriod chat application with GCP Firebase backend. This application supported single and group conversations, with real time firebase message delivery. This application was a lite application which could be easily integrated into other android applications.
  4. Cognitive Computing – Idea Labs, Nirma University
    Created a system that calibrated brain waves from an EEG headset to determine the recall ability of a person while performing a forced choice memory test on flags images. The recall of images was predicted on time series data using LSTM RNN. The precision of this classifier of this one of a kind test was around 0.7.
  5. Adaptive Video Player
    Developed a web application that recognized student behavior while studying from video lectures. Data is streamed from an EEG headset, eye position from a camera (using OpenCV), and heart rate from a sensor to NoSQL database. And, Based on this data, an algorithm was designed to calibrate the student’s attention level and using javascript the playback speed of the web video player was adapted.
Past Publications 📖
  1. Title: Emotion recognition from sensory and bio-signals: A survey
    Conference: Proceedings of the 2nd International Conference on Data Engineering and Communication Technology: ICDECT 2017 link
  2. Title: Necessary Precautions in Cognitive Tutoring System
    Conference: Advances in Intelligent Systems and Computing link