My areas of interest include Probabilistic Modeling for Natural Language Processing(NLP) and CV Systems & Data Analytics with a focus on the domains of Machine Learning & AI Engineering, Image / Video Analysis
I am a part of Maritime Graphics team at Fraunhofer IGD, Rostock. We are working on problems related to underwater computer vision and other related areas.
I am part of a team working on making Image Segmentation models lightweight for use in mobile devices. We mainly work with efficient architectures for Convolutional Neural Networks, network pruning etc.
I was part of ML team at Haptik. I was working with DL Based text classification, supervised sentence similarity and sequence classification for NER and dialogue management.
At jubi.ai I worked on their chatbot development platform. I developed modules like text ranking, text classification for intent recognition and named entity recognition.
I am pursuing masters in Computational Engineering at FAU. My application field is Information Technology - Digital signal processing. The course is a combination of Mathenatics, Computer Science and Digital Signal Processing. Some of the important courses I have studied/studying listed below.
I am pursuing masters in Computational Science at USI. I am part of a Double-Degree Programme at FAU and USI. Courses I took at USI
My undregrad was in Electronics Engineering from KNIT, an intitute of Uttar Pradesh Goverment. Below is brief list of the subjects I studied.
Exploring contributions of bio-signals and facial expressions for multi-modal emotion recognition and developing fusion strategies using deep learning.
Thesis(Low-Res) Thesis(High-Res)Studying and analyzing latest developments in efficient architectures of Image Segmentation. Project: Create UNet like architectures based on more efficient and sparse architectures like Depth-Wise separable and grouped convolutions. Goal: Experiment with Atrous Convolutions, Depth-Wise Separable Convolutions, Channel Shuffle, Bottleneck Layers etc to reduce FLOPs while maintaining accuracy of the model.
Oracle was built to make data centric ML development more efficient. It helps data analysts see the problems with the data for intent classification task at Haptik. Oracle Insights such as "Points of intent(class) overlap" - by analysing the ditribution of prediction probabilities, helped data analysts modify the dataset effectively leading to more accurate results.
IVA makes a decision on what the user wants by detecting Intent and Entities from the incoming user-query. I worked on the NER module that Haptik's IVAs use. Multiple experiments were done using BERT encoder with a CRF head on SNIPS custom-intents dataset. Considering the requirement that the system should be able to detect different types of entites from small, ill-formed sentences we settled with CRF classifer on top of GloVe embeddings. The main problem we faced was that since most entities in our IVAs used be OOVs, in cases when in-vocabulary word(s) appeared in an entity it wasn't recognized. After a lot of analysis we found the problem was in GloVe vectors where due to very strong representation, the word was almost never classified as an entity. We solved the problem by implementing a dropout like scheme for this embedding vector.
Domain Classification/Tagging based on entity embeddings. Parsed 400GB freebase data to extract entities and their definitions/descriptions. Finetuned GloVe vectors to include these entities. Trained a text classification system using Flair and pooled embeddings to pool BERT outputs with custom GloVe vectors
LSTM and Bilinear Layer model to interpret semantic similarity between two short excerpts of text( messages in a chat) It consisted of an LSTM encoder followed by a bilinear-layer(tensor layer). The model was trained on chat data. Pretrained glove(twitter) embeddings were first fine-tuned on 200MB of textual chat-like data For eg. youtube comments, twitter conversations, daily-dialog, etc. and then used while training the model. The model was able to differentiate between adversarial sentences like “I love you” and “I hate you”; this is one of the cases where existing(at that time) sentence encoders fail. This architecture was used to deploy models for Smalltalk/chit chat for 10 languages
View ProjectVariational autoencoder was used to build an architecture proposed by Gupta et al, A Deep Generative Framework for Paraphrase Generation. It was trained using quora similar questions dataset.
View ProjectWe are trying to help content creators in India monetize their content, either with private delivery channels or finding them the right distributor. Apart from leading and managing a team of 4 engineers, for frontend and UI development. I worked on the backend myself developing a Content Management Platform using Django framework and a video encoding service using ffmpeg, bento tools. Deployed the entire system on AWS ECS and AWS Batch(for encoding jobs).
View ProjectI co-built the IVA development platform at jubi.ai. The system consisted of modules like text ranking, intent and entity detection.