Sayan Nath

Kalinga Institute of Industrial Technology

CGPA: 9.02

ML Engineer

Roles/Responsibilities
• TBA

Google Summer of Code Student

Roles/Responsibilities
• Understand the Tensorflow Lite Task Library.
• Implemented CameraX and remove the usage of fragments with the existing Camera2 and Camera API in Object Detection App.
• Implement Support Library with TensorFlow Lite Interpreter.
• Implemented Data Binding.
• Modified the Test Code in Android.
• Adding a Bounding Box Function in tflite-support Library.

Backend Developer

Roles/Responsibilities
• Responsible for handling the Backend and Scalability.
• Implement many alternative solution for a problem.
• Maintaining Design Docs and API documentation.
• Managing AWS Server as well as MongoDB Atlas and Mongo Realm.
• Implemented and Deployed Image Moderation API with Docker.

Data Science Intern

Roles/Responsibilities
• Worked on Optical Character Recognition(OCR) with Tensorflow, Keras and TFLite. Integrated in Flutter App as well.
• Created the backend using Nodejs, Express, Redis, Socket.io, MongoDB and maintained AWS Server with EC2 instances and S3 Bucket.
• Scaled the complete backend with horizontal scaling to increase efficiency by 70%
• Mentored 2 engineers to deploy app in Play Store and App Store. Also got awarded for youngest talent

Machine Learning Engineer

Roles/Responsibilities
• Build many Computer Vision Models as per customer requirement. Optimised the models with various optimisation technique like Quantisation, TensorRT Weight Pruning and Weight Clustering. Reduced the inference time of the model from 800ms to 250ms.
• Played a major role for developing Ikshana AI. Worked with TensorFlow Object Detection API, TensorFlow JS and TensorFlow Lite to build the model and deploy it in edge devices.
• Worked with big companies like Mahindra and PropDial.
• Identified Software and DB Architecture based as per customer requirements, performed version control and refining product.
• Created the backend using Nodejs, Express, MongoDB, MySQL and maintained Digital Ocean Server with Droplet and Storage bucket.

Flutter Developer Intern

Roles/Responsibilities
• Added new features to the exsiting app.
• Made an app after understanding business requirements.

President

I am President at CodeChef KIIT. In CodeChef KIIT I take core decisions and manages my team to conduct events, sessions and contest in KIIT University

Lead

Currently leading the Google GDSC KIIT team. Managing the team and organise the events.

Machine Learning Lead

I am in-charge of Machine Learning Team. In IoT Lab KiiT I do projects along with my team which are related to Machine Learning and Deep Learning. I do take core team decisions with my team.

Sanus

It is an end to end Deep Learning Project with a Flutter App. Simple Website (with Html5, Css3, JS and Flask) to detect diseases like Heart Disease, Pneumonia, Skin Cancer, Brain Tumor, and Malaria Detection by analysing the MRI, CT Scan, Blood Cell Images and taking important data to predict the output.

Github Link

DCGAN-TensorFlow2

Generating Abstract Images using DCGANs

Github Link

American Sign Language Detection App

A realtime image detection app which helps to detect American Sign Language. It handles upto 29 classes. Used MobileNetV2 to train the images. It is deployed in smartphones using TF-Lite.

Github Link

Face Authentication App

Face Authentication with TF-Line. Model architecture which is used is FaceNet.

Github Link

Low Light Image Enhancement App

MIRNet model is used to enhance a low-light image. Implemented MIRNet TF-Lite Model with Flutter App.

Github Link

Spyd App

An android app that detects whether the signal of your app is weak or strong. If the signal in your area is poor or weak, the current location of the user gets stored in the database. According to the network provider, the data is saved in the Firebase Realtime-Database.

Github Link

ZERO

Developed two Android and two iOS for both customer and vendor and one Android App for the delivery personell. The workflow of the software is pretty smooth and client is also happy with the deliverables.

Work Link

Quiz Adda

Developed one Android App for users to buy subscription and play quiz. Admin portal to manage quiz and give rewards to the partcipants.

Work Link

PropDial

Developed an internal app for real estate agents to manage their properties, daily task and to log their employee's attendance.

Work Link

Rahat

Presented in Project Innovations in Distributed Computing and Internet Technology Proceedings of 10th PIC 2021.

Paper Link

How to make Computer Vision Project with TensorFlow & Keras

In This Session, I have talked about how to make Computer Vision projects with TensorFlow and Keras. Discussed about TensorFlow Ecosystem from Data Collection to Model Serving. Also discussed about best practices.

Deck

Session Link

Introduction to TF-Lite

In This Session, I have talked about TFLite. TF Lite models are lightweight models, production-ready and cross-platform frameworks for deploying ML models that are used to get inferences on edge devices like mobile phones and microcontrollers.

Deck

Session Link

Info Session: Solution Challenge 2021

DSC KIIT has conducted an info session for all interested student developers to get an idea about the event, timeline, steps and any doubts you might have for DSC Solution Challenge 2021.

Deck

Session Link

Transfer Learning with Keras

Without transfer learning, it would've been very difficult for beginners to dive into ML. But now, with just a fraction of resources you can make really good models. Therefore DSC KIIT brings you a session on Transfer Learning with Keras in which we'll build an amazing project from scratch and learn necessary concepts along the way.

Deck

Session Link

Deep Learning with Tensorflow 2.0

In this session I talked about making simple Convolution Neural Network with Tensorflow 2.0 and demonstrated a project on Malaria cell detection.

Deck

Session Link

Dart Complete Course

A full tutorial series on the programming language of Dart. I have taught basic, intermediate as well as advance concepts of Dart.

Course Playlist

Build Alexa Skill

Amazon Alexa is a virtual assistant AI by Amazon.It works on amazon echo dot and many more home automation devices.It is also capable of making to-do lists,set alarms play audiobooks,etc. We have built a DSCKIIT skill.It will help you know about various facts of DSCKIIT and also other developer student clubs running all over India.

Session Link

Explore ML Workshop

Conducted Explore ML Workshop in my college. Took an Intermediate Session in this workshop. Shared roadmaps and solved doubts of the students.

Deck

Photo Link

Image Classification using BigTransfer(BiT)

BigTransfer (BiT) State-of-the-art transfer learning for image classification.

Blog Link

CutMix data augmentation for image classification

Data augmentation with CutMix for image classification on CIFAR-10.

Blog Link

How to cache JSON data in Redis

In this article, I have talked about how you can cache JSON data in Redis using NodeJS and MySQL.

Blog Link

Weight Pruning with Keras

In this article, I have talked about weight pruning with Keras which is one of the methods to optimise your model.

Blog Link

Quantisation of Models

In this article, we will understand how quantisation works with TF-Lite Models.

Blog Link

Introduction to TF Lite

In this article, I have talked about saving tensorflow models in TF Lite format.

Blog Link

Theme Changer using Provider in Flutter

In this article, I have made a Flutter app where we can change the theme of the app by using Provider and Shared Preferences.

Blog Link