Shreyas Bapat

Software Engineer at Shuttl

How does this work?
About

Ever since I was a little kid, I was passionate about all the latest tech in the market. Astronomy and Space Technology always used to include the latest and bleeding-edge tech. I first saw Linux being used in one of the high-end observatories in ARIES Nainital, and then saw Himalayan Chandra Telescope being controlled by a remote operations center in Bengaluru. These experiences made me curious about automation and the effects it can have on mankind. I joined IIT Mandi for my Bachelor of Technology program with majors in Electrical Engineering and then found out the vast amount of information that was there in this field. I took up courses relating to Distributed Systems, Computer Organisation, Pattern Recognition and some courses on Algorithms and Deep Learning. I took up a very challenging project as my Major Technical Project titled "Computational VLBI Image Reconstruction" which involves reconstruction of images acquired from Event Horizon Telescope. During my summer breaks, I found out that there is an amazing world of Open Software on platforms like GitHub and GitLab. I started contributions to important packages in Astronomy and Astrophysics. And learned various Software Development Practices, Continuous Integration and Continuous Deployment, Technical Documentation and Performance Optimisations, etc. I eventually developed my own Python Package on General Relativity named "EinsteinPy" through which I learned Packaging, Software Lifecycle, etc. I ended up becoming a SOCIS and GSoC Mentor for this. I have got the privilege of working at Siemens Research in India where I worked on Deep Learning Applications on style transfer etc. I explored a completely new field of Active Learning. As I went for my second internship in the same company, I worked on Validation of Deep Neural Networks and coverage guided fuzzing, etc. Now I work in a cloud computing startup called Semut, where I take care of their distributed storage, cluster efficiency, and production Kubernetes cluster as well. I am also actively working on my own startup called Furrble, where we are actively working on a Flutter Android/iOS App, a scalable backend and all the infrastructure required. I also am learning a lot of management in the process. So, now in the industry, the passion remains in me. There is a lot to learn, a lot of problems to solve, many more theories like General Relativity to be conquered through software and a lot of lives to be led to betterment because of software.

Ask me about
Data Science and Machine Learning Platforms
Red Hat Ceph Storage
AI powered customer data and intelligence platform
Kubernetes
Work experience
logo

Apr 2021 - Present

Shuttl

Software Engineer

Jul 2020 - Mar 2021

Semut.io

SDE 1 - Cloud Computing R&D

Worked extensively on cloud technologies like Kubernetes, Kata, Docker, AWS, Ceph, Rook etc.

logo

Jun 2019 - Jul 2019

Siemens Technology India

Software Research Intern

• Filed a patent on the results of the Style Transfer Deep Learning algorithm. • Created an application with user interface for finding coverage of a neural network and its class activation maps. • Worked on benchmarking the results produced by CycleGAN and Multimodal Unsupervised Image-to-Image Translation method, and explored some aspects of Explainable Artificial Intelligence (AI), testing and validating the neural network.

logo

Dec 2018 - Feb 2019

Siemens Technology India

Software Research Intern

• Used generative models for test data generation. Exploring active learning for automatic data labelling. Explored and evaluated the algorithms for text to video generation and created a report for it. • Created a software that can transform the weather in a photo (e.g. sunny to rainy), which was achieved by Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks.

Aug 2018 - Nov 2018

ANKAM

Deep Learning Intern

• Used transfer learning, residual learning to classify images of human eyes using ResNet50 for Diabetic Retinopathy Detection. Achieved an accuracy of 86%. • Created a web interface in flask and a RESTful API for analysing the retina images to be used by different hospitals in Delhi and Bengaluru Region.

Talk to Shreyas

@ Copyright 2020 OfficeHours Technologies Co.