Projects
DDOS Attack Detection (Machine Learning)
Developed a lightweight classification algorithm with an accuracy of 95% to classify incoming network packets as Benign or Malicious.
Skills: Support Vector Machine, Decision Tree, Naive Bayes, Adaboost, Ensemble, XGBoost, Keras, sklearn, K-Means, Machine Learning.
Clinical Text Extraction
Built NLP Pipeline for stemming, lemmatization, tagging, and Named Entity Recognition to analyze patient health records to extract the diseases, dosage, and treatment information.
Skills: Data Mining, LSTM, GRUs Seq2Seq, NLP, Tensorflow, Parts-of-Speech Tagging.
Sentiment Analysis on Movie Reviews
Implemented sentiment analysis algorithms such as logistic regression, Naive Bayes, and BERT for classifying a sentence as positive or negative sentiment.
Skills: NLP, Logistic Regression, KNN, Locality hash, Transformers, Encoder-Decoders, T5 + BERT, Word Embeddings, Siamese Networks.
Vulnerabilities in Neural Networks
Achieved 16% improvement in energy demand prediction by analyzing model failures due to noise perturbation in input data through deep and shallow neural networks.
Skills: Neural Networks, Deep and Shallow Networks, Preventing curse of dimensions, Pytorch.
Student Behaviour Analysis
Conducted a survey to analyze the student behavior for assignment submission.
Skills: Hypothesis Testing, T-test, ANOVA, Regression, Normality test, Gaussian Distribution.
Security and Energy Aware Task Scheduling in MPSoCs
Employed techniques from FreeRTOS to design a dual optimal solution which enhances security and energy utilization in heterogeneous MPSoCs, improving Task Schedulability by 8%.
Skills: Optimization, Heuristic Development, FreeRTOS, Novelty.