Shanu Vashishtha

I am a Deep Learning Engineer at Rain Neuromorphics. I completed my graduate and undergraduate degree at the University of Massachusetts Amherst and IIT Kanpur respectively.

Email  /  LinkedIn  /  CV  /  Medium

I am interested in applying deep learning techniques to solve problems in the domain of computer vision and natural language processing.

In my current role, I have focused on training neural network based models for a novel analog neuromorphic hardware.

profile photo
Work

I worked full-time as a Senior Engineer at Honeywell Technology Solutions, Bangalore. I have also done internships at Rain Neuromorphics, Computer Vision Center (UAB) and Inria Sophia Antipolis.

Projects

Sparse Linear Library [Code]

Pytorch package that allows users to create extremely wide and sparse linear layers efficiently. Supports unstructured sparsity, static sparsity as well as a dynamic growth and pruning strategy

Augmented Transformer for Machine Translation [Report]

Combine Part-of-Speech, predicate and word embeddings to incorporate linguistic information, experiemnt with pre-trained contextualized BERT embeddings and Syntactically informed self-attention layer

Classifying Animal Species in Camera Trap Images [Report]

Use feature extraction, finetuning, mixup and bounding box proposals to classify animal species in camera trap images

Visualization Tool for Large Scale Comments [Report] [Code]

Define, Empathize, Ideate, build prototype and test for creating a tool used to gather relevant product information in limited time

Using Microsoft Azure as a Machine Learning Service [Report] [Code]

Produce user-friendly, end-to-end data science notebook tutorials that are accessible to general data science audience

Neural networks to classify aircraft images [Report]

Bilinear CNNs for fine-grained visual classification of manufacturer and variant category of aircraft images

Miscellaneous

To Drop or Not [Article]

‘Tipsy’ teens blame it on peer pressure [Article]

The Semester that was: GRC’s session with professors [Article]

Engineering Sciences: The Real Picture [Article]