NIKHIL VAIDHYANATHAN

Dynamic and motivated Computer Science graduate with experience in working in the field of large language models, machine learning and data science. Proficient in Python, R and SQL. Adept at collaborating with cross-functional teams to form solutions. Passionate about continuous learning on current and emerging technologies in the field.

• Data Analytics Professional Certificate, Google
• Agile Meets Design Thinking, University of Virginia
• Entrepreneurship Essentials, NPTEL

• Malaria Species Detection : Tackled the challenge of predicting malaria species by conducting a comparative study using 10-12 machine learning models, including Logistic Regression, Naive Bayes, ANN, KNN, Decision Trees, ADABoost, SVM, and XGBoost. Utilizing Python and Jupyter Notebooks, we evaluated each model's effectiveness on our dataset, providing valuable insights for future advancements in malaria detection.

• Identification and Classification of Toxic Comments using Deep Learning : Developed a deep learning model for toxic comment detection in social media platforms in order to promote a safer online community. The model is built using a bi-directional LSTM Layer and three dense layers with ReLU activation and different weights. The toxicity is classified according to their intensities and flagged for further review

• Machine Translation between Tamil and English using Transformer Model's Positional Encoders : The project focuses on the Transformer model, with an encoder-decoder structure used for machine translation. It preprocesses input text by tokenizing and encoding it into vectors, with positional encoding. Trained on a parallel corpus of Tamil and English sentences, we explored how various hyper-parameters affected performance. Results indicated that increasing the number of layers and hidden layer size enhanced the model's performance.