I am pursuing my research master's in Computer Science at DIRO, Université de Montréal with Prof. Noam Aigerman. My graduate research focuses on revamping neural network architectures for 3D geometry and Computer Vision. Currently, I am also doing some cool work on real-time neural rendering of facial animations at Ubisoft La Forge.

My intellectual interests lie in niches of 3D Computer Vision, Vision-Language Models, Neural Geometry Processing and Neuro-inspired AI. I also have some experience in programming shaders for engines (HLSL), GPU programming (CUDA) and Quantum Computing (Qiskit).

I did my bachelor's in Computer Science and Engineering at National Institute of Technology Silchar.
I worked on text-conditioned 3D motion generation under the guidance of Dr. Arjun Chandrasekaran, from Micahel Black's group at Perceiving Systems department of Max Planck Institute for Intelligent Systems.
Prior to that, I completed my internship at Amazon as an Applied Scientist Intern. In 2021, I interned at ISRO advised by Dr. Santonu Goswami. I also interned at IIT Patna advised by Dr. Sriparna Saha.
I was also associated with Centre for Natural Language Processing at NIT Silchar, advised by Dr. Partha Pakray for my bachelor's thesis.
Project Lead @ ML Club, NIT Silchar.

I like indulging in open-source shared tasks and hackathons where intelligent systems can be utilized to solve problems of humanity. If you wish to collaborate for any such innovative initiative, feel free contact me.

Experience

· Research Developer intern at Ubisoft La Forge

· Research project in collaboration with Dr. Arjun Chandrasekaran, Perceiving Systems, Max Planck Institute for Intelligent Systems

· Applied Scientist intern at Amazon

· Research intern at ISRO (Indian Space Research Organisation)

· Research intern at IIT (Indian Institute of Technology) Patna

· Associated with CNLP (Centre for Natural Language Processing), NIT Silchar

· Qiskit Advocate, IBM Qiskit Advocate Program

· ML Wing Head, Computer Science Society, NIT Silchar

· Moderator, Machine Learning Club, NIT Silchar

· Winner, HackHarvard, MLH Season '22

· Winner, DiCOVA Challenge 2021 (Track-1)

Publications

Wikipedia Current Events Summarization using Particle Swarm Optimization · 2021
Santosh Kumar Mishra, Darsh Kaushik, Sriparna Saha, Pushpak Bhattacharyya
Proceedings of the 18th International Conference on Natural Language Processing (ICON), 2021
International Conference on Natural Language Processing (ICON), 2021
[ACL Anthology] [PDF]


Particle Swarm optimized clustering technique for enhancing unsupervised Multi-document Text Summarization.

Improved English to Hindi Multimodal Neural Machine Translation · 2021
SR Laskar, AFUR Khilji, Darsh Kaushik, P Pakray, S Bandyopadhyay
Proceedings of the 8th Workshop on Asian Translation (WAT2021), 155-160
International Joint Conference on Natural Language Processing (IJCNLP), 2021
[ACL Anthology] [PDF]


Leveraging visual features to improve low-resource pair translation.

COVID-19 Diagnosis from Cough Acoustics using ConvNets and Data Augmentation · 2021
SK Mahanta, Darsh Kaushik, S Jain, H Van Truong, K Guha
2021 First International Conference on Advances in Computing and Future Communication Technologies (ICACFCT)
[ IEEE Xplore ] [PDF]


Predicting COVID-19 infection from cough acoustics.

CNLP-NITS@ LongSumm 2021: TextRank Variant for Generating Long Summaries · 2021
Darsh Kaushik, AFUR Khilji, U Sinha, P Pakray
Proceedings of the Second Workshop on Scholarly Document Processing, 103-109
North American Chapter of the Association for Computational Linguistics, 2021
[ACL Anthology] [PDF]


Generating longer summaries to extract salient information in scientific documents for LongSumm 2021 shared task.

EnKhCorp1.0: An English–Khasi Corpus · 2021
SR Laskar, AFUR Khilji, Darsh Kaushik, P Pakray, S Bandyopadhyay
Proceedings of the 4th Workshop on Technologies for MT of Low Resource Languages (LoResMT2021)
Machine Translation Summit 2021
[ACL Anthology] [PDF]


Developed a corpus and baseline machine translation systems for low resource English–Khasi pair.

Past Projects

Flomeomorphisms · 2024-25
Neural Geometry Processing, 3D Shape Analysis, Bijective Fields
Darsh Kaushik
, Noam Aigerman
Manuscript in preparation


First neural architecture for fast, accurate and bijective flows on 3D meshe to learn surface maps and shape correspondences.

ASyMov · 2022-23
Multimodal Deep Learning, Latent Variable Models, 3D Computer Vision
Darsh Kaushik,
Arjun Chandrasekaran


Using temporally discretized 3D motion sequences to enhance human motion generation from textual prompts.

Cough It · 2021
Audio Signal Processing, Convolutional Neural Networks, Data Augmentation
Darsh Kaushik, Jyotimoy Kashyap, Gaurav Das
Winner of HackHarvard 2021
Highest ranked model at DiCOVA Challenge 2021
[ Code ] [Report]


Scalable mobile application for simple, efficient and inexpensive COVID-19 diagnosis from cough audio samples.

GradCAM · 2020
ML Interpretability, Computer Vision, Transfer Learning
Darsh Kaushik, Kavii Suri, Asis Kumar Roy
[ Code ] [Website]


Deployed model on the web for skin lesion classification and interpretation using GradCAM.

Other Earths · 2020
Signal Processing, Time series, AWS
Darsh Kaushik
AWS Machine Learning Scholarship Program · Capstone Project
[ Code ] [Report]


Deployed a classifier on AWS to identify stars with exoplanets using the transit method on time-series of flux signal from Kepler K2 mission by NASA.