• Simulating molecular dynamics trajectories in a parallelized mode using Ray.
• Analyze the obtained trajectories and reduce the dimensions of the trajectories using Auto-encoder.
• Choose the new restart conformations for the next set of molecular dynamics trajectories using Curiosity Through Random Network Distillation (RND).

Technologies: Ray, Python, OpenMM, and Git.
I wrote a scientific article, around the results of my master thesis, entitled "Effectiveness of parallelism in the adaptive sampling process to enhance protein conformational sampling" for PeerJ Journal.

Analyze Johns Hopkins University data using unsupervised models

Analyzed Johns Hopkins University data using K-means, Hierarchical classification and Principal component analysis (PCA).
In this project, I learned the importance of data quality which really describes reality to get good and meaningful results.
Technologies: Pandas, NumPy, Matplotlib, Scikit-Learn and Python.

Fake News Detection

Develop a machine learning algorithm to identify when an article might be fake news using Natural language processing (NLP), support vector machine (SVM).
Technologies: NLTK and Python.