I train machines to make them smarter than me. I fail continuously.
Glancing across the flood of online advertisements on Facebook, in July 2015, ‘Machine Learning’ (ML) caught my attention, and I was eager to learn more. Since then I've never looked back!
I am currently pursuing my Masters in CS from the University of Utah with a focus on ML and DL. Before starting my masters, I worked as a Sr. Data Scientist for two years at Algoscale Technologies. I also worked as a Deep Learning Intern at Intel Corporation in Summer 2020.
I always try to make a change in the world in any and every field possible. I have blended my CS skills with marketing, healthcare, finance, etc. to bring out the best of both worlds. Please look at my projects to learn more.
PORTFOLIO OF WORK
If you have any suggestions or would like to collaborate with me on the next world-changing project.
I am just a click away!
DEEP LEARNING ENGINEER
May 2020 - Aug 2020
Optimized Facebook’s RetinaNet model with 2.1x improvement over native TensorFlow version using parallel programming in both C++ and Python
Worked on improving training performance of Google’s TensorFlow for Intel scalable CPUs using MKL library on models like ResNet50, RetinaNet, MaskRCNN, Transformer and BERT-Large
Algoscale Technologies Inc.
Jul 2017 - July 2019
Developed an Auto-Deep Learning pipeline that allowed non-technical users to train machine learning models using Keras (Python) with features such as memory management, result storage, and remote access
Built an automated Cryptocurrency Trading Bot on Bittrex exchange with a 10% profit on each transaction based on historical features and financial indicators using XGBoost in Python
Built a recommendation model for medical drug trials based on cosine similarity of word2vec features ensuring 95% accuracy, parsing over 250k XMLs stored in Elasticsearch, using Python
MACHINE LEARNING RESEARCHER
Jan 2017 - June 2017
Collaborated with AIIMS, Delhi, India to develop a mobile platform for Diagnosis of Cervical Cancer by classifying cervical images using ResNet18 into severely infected, infected and normal, in Keras, helping around 60%cervical cancer patients in Orissa, India