Hi, my name is

Build a Deep Learning based model using Recurrent Neural Network or/and GRU/LSTM for Digital Twin of the Cryogenic Energy Storage System integrated with a Matlab simulator.

The above process required exerimentation wih various deep learning models and architecture and to make observation whether the model can make prediction successfully or not after only seeing small amount of data (few-shot) and without any prior domain knowledge.

The model also made decisions and helped the operator to take operational decision on the plant for error free working of the energy storage system.

During this internship I worked on building a recommendation system using machine learning and the LightFM Framework provided by Lyst. The recommendaion sysem will give suggesions to users of the Tweencraft Animation App. The problem of coldstart was handled by making use of a hybrid app.

I had also built REST APIs using NodeJS in the backend and MongoDB for accessing data. These APIs were used to integrate the recommendation model with the application.

Built complex queries using SQL and BigQuery (on GCP) to fetch meaningful data and make interactive visualizatiions with Google Data Studio so tha actionable business decisions can be made on that.

Conducted workshop for Open Source Awareness like Python and Git workshop to provide sudents with the basic technology knowledge and help them get started with contributing to the Open Source Projects.

Conduct KWoC which is a small scale version of GSoC implemented on the local level to promote contributions open source projects. We encourage and approach both students and mentors to participate in the contribution marathon

Also conducted Open Source Summit as a part of Ksitij (Techno-Managaement fest IIT Kharagpur) where a bunche of interesting workshops including workshops related to Ethical Hacking and Golang were conducted, and the workshops had participation from students even from other college (apart from IIT Kharagpur)

Building a deep learning model and training it using meta learning algorithms to observe whether we can build a model that is efficient enough to solve puzzles only given the visual puzzle and he naural language rules given as a list. For this we start building model on Abstraction and Reasoning Challenge hat was conducted on kaggle.

Following this we will use the same idea to build a model for chatbot application, that will be learning on the fly and will be able learn about different domain speciific tasks from the recent context. Also when the chatbot fails it should be able to generate appropriate feedback question ha will help o improve is performance. For this we are making use of the MetaLWOz dataset.

As the part of this society we promote Sport programming (Competitive programming). Students are encouraged to participate in various competitions where we provide a platform to discuss and learn new concepts from each oher while practicing.

We also conduct various contests that involve challenging questions around the year to keep sudents engaged in competitive programming and also help get prepared for the internship tests by practicing through various practice tests.

Build problems which test algorithmic and data structure knowledge from varied domains like dynamic programming and greedy approaches.

Test the problems already set for the upcoming contests organised by the organisation

I was also given the responsibiity of Conent Creation & Doubt Clearance while understanding the need of the team in erms of quality and the dimesion of the problems required.