• Experienced in data analysis, data pre-processing, model building/optimization, and deployment.
• Passionate about using data analysis and finding the key patterns and recurring frequencies in the data which can help revolutionize the way we look at business problems. A strong believer in the concept of data cleaning as it can directly affect the performance of the model which can easily be overlooked. Data cleaning helps in building a strong and sustainable model.
• Ample knowledge of Machine Learning from basic algorithms like Naive Bayes, logistic regression to complex algorithms like bagging and boosting. Well-versed with Keras and Tensorflow to use deep learning algorithms from simple MLPs to complex algorithms like VGG, Resnet, LSTMs, Bert, etc.
• Able to formulate algorithms from scratch based on new research papers which are not yet available in the well-known libraries.

PROJECT DESCRIPTIONS:

Microsoft Malware Classification

● The goal of the project is to classify the type of Malware File using the Machine Learning technique.

● Performed Univariate Analysis on size of .asm & .bytes files as well as on some of the opcodes of asm files.

● Tools & Technologies: sklearn, XGboost, XGBClassifier, CalibratedClassifierCV, RandomForestClassifier, RandomSearchCV, SelectFromModel, Dask, log_loss, confusion_matrix, train_test_split, codecs, multiprocessing, tqdm, Pandas, numpy, pickle, imageio, seaborn Custom Ensemble

● Custom Ensemble using Decision Trees for Jane Street Market Prediction problem.

● Tools & Technologies: random, math, pickle, numpy, pandas, clone, f1_score, DecisionTreeClassifier CalibratedClassifierCV, RandomizedSearchCV, GroupTimeSeriesSplit.

Face Recognition

● Using various techniques to count and detect the number of Faces from the image, and then perform Face Recognition using Deep Learning Techniques such as Face Net (Google pre-trained model) and Transfer Learning.

● Tools and Technologies: OpenCV, mtcnn, and TensorFlow. Legal document’s classification using ML & DL

● Natural Language Processing (NLP) based project, using OCR tool user can upload legal documents in Image (JPG, JPEG, PNG), Pdf and Docx format, then text classifiers can automatically analyze text and then assign a set of pre-defined tags or categories based on its content. Text Classification we have proposed Machine Learning and Deep Learning approaches. For the Machine Learning approach KNN, SVC & Naïve Bayes are used while the BERT model is used for the Deep Learning approach.

● Tools and Technologies: Pandas, NumPy, OpenCV, Tensorflow, scikitlearn, Django, Nodejs.