iamtejas

"The world is full of magical things, patiently waiting for our senses to grow sharper."

About Me

Hello, I am Tejas Vaidhya (you can also call me Dev),
I am an Undergrad Researcher, Passionate Opensource developer, and a tech enthusiast. I am currently in my third year of Undergraduate studies, pursuing my major in Bachelor of Architecture and Planning and Minor in Mathematics and Computing at IIT Kharagpur.
My research interest lies in Computer vision and Natural language processing. Currently, I am researching on Probing techniques for Transformer based model.
I love Programming, Designing, and Planning (Afterall I am an architecture major). I also enjoy watching Anime, reading non-fiction, film making, cooking, and interacting with new people. I am a foodie. In my free time, I like to explore food (you can often find me in some of the eateries of KGP). I have always adored technology and am very much fascinated by how it affects the life of people.

Experience

Student Developer

Google Summer of Code @julialang,

Sep 2017 — present

Working on framwork for Statistical and Deeplearing based Language model in julia.

Implemented Statistical Language model framework to develop traditional NLP based model in julia.
Framework for Alite BERT in julia for reseracher, students and Industries
Julia Wrapper for Sentencepiece library

NLPTransformerLanguage modeljulia

Student Researcher

CERN, Hong kong

march 2020 — Present

I am involved in the ongoing research on the track classification for identifying tau-leptons from Boson decay. I am working on developing deep learning based algorithms to efficiently classify between the background noise and tau-jets.

DeepLearningLHCTau-leptons Decay

UG researcher

Centre for Artificial Intelligence,

Aug 2019 — March 2020

Developed a bidirectional autoencoder based model which uses the information extracted from context with out of reference knowledge to get the softmax probabilities over all the possible answers to a question
Extracted the edges based relations from commonsense corporas like ConceptNet and NELL
Working on making commonsense inclusive word vectors using the relations extracted like ConceptNet Numberbatch

Natural Language ProcessingMachine Learning

details of projects can be found here