Subhani Shaik

mine

Education

Indraprastha Institute of Information Technology (IIIT) - Delhi M.Tech CSE 2018 - 2020	CGPA: 8.04
Rajiv Gandhi University of Knowledge Technologies (RGUKT-Nuzvid), Andhra Pradesh B.Tech CSE 2013 - 2017	CGPA: 8.8
Rajiv Gandhi University of Knowledge Technologies (RGUKT-Nuzvid), Andhra Pradesh Pre-University Course 2011 - 2013	CGPA: 8.3
Kondaveedu Public School, Andhra Pradesh 10^th Standard 2010 - 2011	Percentage: 95.6

FaceMark
Guide: Dr. Pushpendra Singh, IIITD				(Sep,19 – Oct,19)
The goal of the project was to Mark the attendance using Face Recognition and generate the report. We developed an Android-based mobile application to identify the student faces from a camera captured image of the classroom using MTCNN and FaceNet based face recognition model.
Dataset: Collected from students \| Code

* Plant Disease Detection
Guide: Dr. Richa Singh, IIITD				(Oct,19 – Nov,19)
In this research project, we analyzed the dataset of plants and we experimented on various CNN based models -AlexNet, ResNet on variations of dataset and developed a Multi-Tasking deep learning model to detect the specific disease and compared our results with then state-of-the-art models.
Dataset: Plant Village image dataset (PlantVillage.org) \| Code

Image Dehazing
Guide: Dr. A V Subramanyam, IIITD				(Oct,19 – Nov,19)
In this research project, we analyzed the dataset and developed models to remove haze in the images based on single image using classical Digital image processing algorithms -Dark channel prior algorithm, Histogram equalization and matching techniques.
Dataset: I-Haze, O-Haze dataset (NTIRE 2018 challenge) \| Code

SVM and SVM-Ensembles for Breast Cancer Prediction
Guide: Dr. G.P.S Raghava, IIITD				(Oct,19 – Nov,19)
In this project, we analyzed the dataset and developed SVMs(using different kernels and an advanced SVM variant NuSVM), their ensembles. We compared our results with existing journals and observed that our model outperformed existing implementations.
Dataset: Breast Cancer Wisconsin dataset, ACM SIGKDD Cup 2008 challenge dataset. \| Code

* Image Super-Resolution using GANs
Guide: Dr. Saket Anand, IIITD				(Feb,19 – Apr,19)
In this research project, we experimented on AutoEncoders and developed various models - Denoising AE, ESRGAN to Super-resolve low-resolution images to high-resolution images(2x, 4x) and compared the results.
Dataset: DIV2K dataset \| Code

* Extracting Factual and Non-Factual data from News-Articles
Guide: Dr. Tanmoy Chakrabothy, IIITD				(Feb,19 – Apr,19)
In this research project, goal was to develop an extractive model that retrieves factual statements from the news-articles containing factual and non-factual statements like the author’s opinions, predictions, and inferences from facts etc. using Learning techniques. Achieved F1 ~ 95%.
Dataset: MPQA dataset \| Code

Heart Disease Prediction
Guide: Dr. G.P.S Raghava, IIITD				(Mar,19 – Mar,19)
In this project, we analyzed the dataset and developed various classical ML models -Logistic regression, SVM, Naive Bayes, KNN, Decision Tree, Random Forest and MLP and compared our results with existing implementations.
Dataset: StatLog Heart Disease dataset \| Code

Breaking the CAPTCHA
Guide: Dr. Mayank Vatsa, IIITD				(Aug,18 – Dec,18)
In this project, we developed models to decode and recognize the characters in distorted multiple characters CAPTCHA images using SVM’s, K-means algorithms and CNN’s. Achieved F1 ~ 99% for our best model.
Dataset: Generated using ‘captcha’ python-library. \| Code

Sentiment Analysis to classify Abusive comments
Guide: Dr. Saket Anand, IIITD				(Aug,18 – Dec,18)
In this project, we explored a few classical Machine Learning models -Naive Bayes,SVm and Deep Learning models -RNN’s, LSTM’s to detect and classify the abusive comments into specified categories. Achieved F1 ~ 85%.
Dataset: Wikipedia detox and Twitter dataset \| Code

Employee Project Tracking Tool
Guide: Pranideep Kona, Alacriti InfoSystems.				(Jul,17 – Jul,17)
The project goal was to assign and track the project workflow to an employee based on a set of conditions using Java and RestFul web services.
Link: Code

Deep Learning and Machine Learning

Image Style Transfer using CycleGAN.
Scene text Recognition using Convolutional Recurrent Neural Network(C-RNN).
Variational AutoEncoders(VAE) on MNIST and MultiPIE Face Datasets.
Image Classification using ConvNets - AlexNet, ResNet(Transfer Learning on CIFAR10 Dataset).
Object Detection Models - YOLOv3(on WildLife Dataset), Faster R-CNN(on Bosch TLD Dataset).
Speech Recognition using Sequential Models- RNN, LSTM(Audio Dataset).
Neural Network implementation(from Scratch on MNIST Dataset).
Logistic Regression(from Scratch on MNIST Dataset).
SVM(from Scratch on MNIST Dataset).
Linear Regression(from Scratch on Boston Housing Dataset).

Artificial Intelligence

Evolutionary Algorithms: ACO(Ant Colony Optimisation) on TSP problem, Genetic and Memetic Algorithms on Time table scheduling problem.
Searching Algorithms: A*, IDA*, BFS, DFS on N-sliding puzzle and NxN board coloring problem; Mini-max, Alpha-Beta pruning on Tic-Tac-Toe game.
Basic Q-learning and SARSA implementation on Shortest path in Maze.

NLP and Information Retrieval

Languages: Java, Python, C

Web Technologies: Java Script, JQuery, HTML

Tools and Technologies: Pytorch, Scikit-learn, Keras, Open-CV, RESTful-WebServices, SQL, Git, Android, Maven

Expertise Areas: Algorithms, Machine Learning, Deep Learning, Software Development

Worked as a Mentor for freshers at Alacriti InfoSystems.	(Feb,17 – Apr,17)
Organizer in Programming club in my college	(Jan,16 – Apr,16)
Representative for my department.	(Oct,15 – Nov,15)

Qualified Graduate Aptitude Test in Engineering (GATE) 2018 with 96 percentile.
1st prize in ProCoders competition in National Level Tech Fest AFOSEC in 2016 and won many other competitions in technical fest.