Built an open-domain, contextual, conversational QA agent using modified DrQA, BERT and PGNet architectures It has a document retriever, neural reader, clarifying system and an answer ranking moduleTrained an attentive stacked biLSTM model having 45% F1 accuracy on QuAC dataset. Used attentive history embeddings for contextImproved the retriever accuracy (% match in top 5 doc) from 11% to 64% by adding previous QA pairs and topic as context in queries to retrieverResearch Project | Team Size: 1 | Jan'19 - Dec'19 | Github Repo Technologies: Python, PyTorch, NLP, Information Retrieval
Built a framework that generates questions, using transformer based uniLM model, on which a given QA system fails. Trained a 2 layer neural network classifier on failed questions (using infersent embeddings) of a given QA system (BiDAF, BERT based) to classify above generated questions as unanswerable/answerable, with 65% accuracy Research Project | Team Size: 3 | Sep'19 - Oct'19 | Github Repo Technologies: Python, PyTorch, ML, NLP
Developed a backup file system, BKPFS, using a stackable file system WRAPFSBKPFS creates a hidden backup directory and multiple backup files per file to backup any modifications to the file. Users can use ioctls to list, view, recover or delete any of the backups. Wrote ~2000 lines of kernel code to implement system calls, ioctl commands, visibility and retention policies. OS Course Project | Team Size: 1 | Feb'19 - Apr'19 | Github Repo Technologies: C, Linux Kernel Programming, Virtual File Systems, Shell Script
Detects Password reuse. Detect the entering of passwords on the wrong website Warns users on clicking links to non-Alexa 10K websites Used Chrome’s runtime and webNavigation API’s. Stored SHA-512 encrypted passwords in chrome storage. System Security Course Project | Team Size: 3 | Mar'19 - Mar'19 | Github Repo Technologies: JavaScript, Chrome API's
Designed and implemented a Hybrid Neural Object Detection Model from 2 famous single shot object detectors, YOLO and SSD, using transfer learning to detect and classify images Computer Vision Course Project | Team Size: 1 | Feb'19 - Feb'19 | Github Repo Technologies: Python, Tensorflow, Computer Vision
Developed multiple classifiers for hand, upper-body, scene and activity detection in Big Bang Theory Video dataset using own implementation of Logistic Regression, SGD, Hard Negative Mining SVM,Kmeans Clustering and existing CNN and RNN implementations (own architectures) in PyTorch.
ML Course Project | Team Size: 1 | Sep'18 - Dec'18 | Github Repo Technologies: Matlab, Python, LibSVM, PyTorch
Do popular songs endure? Developed a statistical model to predict the popularity of popular songs over time i.e. how popular or not-so-popular songs when originally released are doing today?
Data Science Course Project | Team Size: 3 | Sep'18 -Dec'18 | Github Repo Technologies: Python, Jupyter Notebook
Popular Dataset Contributions on Kaggle Bollywood Movie Dataset | Hollywood Music Dataset
Proposed and tested 5 hypotheses (Ex. black cars value blue cars) using KS, Wald’s & Permutation TestsEstimated Craiglist traffic trough time series prediction of no. of ads posted using EWMA & Auto RegressionProb Stats Course Project | Team Size: 1 | Nov'18 - Nov'18 | Github Repo Technologies: Python
Developed a popular web tool for SPOJ (Sphere Online Judge) users to get correct output for a set of input, decided upon by the user, on a SPOJ Problem. 100 unique page views every day.
Personal Project | Team Size: 2 | Dec'13 - Apr'14 | Website | Facebook Page Technologies: PHP, Javascript, Jquery, Twitter Bootstrap
A mini-gcc compiler which can handle simple sorting programs. Developed for preliminary anaysis of all phases of Compiler Design. Compiler Course Project | Team Size: 1 | Jun'14 - Sep'14 | Github Repo
Technologies: C, Flex, Bison