Kashyap Sureshchandra Patel

Tech: Python, Data Structures, DAG, Object-Oriented Design

• Designed and implemented two custom Version Control Systems in Python inspired by Git internals.
• Implementation 1 uses a JSON-based snapshot model with nested dictionary merges for simplicity.
• Implementation 2 mirrors Git's object database with blob, tree, and commit objects stored in a content-addressed DAG.
• Conducted formal complexity analysis comparing both designs across commit speed, checkout speed, space efficiency, and deduplication.

Tech: PostgreSQL, SQLAlchemy, Pandas

• Designed relational database to analyze research collaboration using arXiv dataset.
• Created ER models and normalized tables for authors, papers, categories, and versions.
• Wrote analytical SQL queries and views to study co-authorship networks and publication growth.

Tech: Machine Learning, NumPy, Pandas, BeautifulSoup, Scikit-Learn, LightGBM, XGBoost

• Built an end-to-end ML pipeline predicting product discount rates from 300,000+ scraped e-commerce pages.
• Performed category extraction, text normalization, feature engineering, and EDA.
• Designed ColumnTransformer pipelines with scaling, target encoding, and TF-IDF.
• Tuned Random Forest achieved RMSE 10.73 and R² = 0.69.

Tech: Deep Learning, PyTorch

• Implemented GAN architecture from scratch to understand generator–discriminator adversarial training dynamics.
• Designed fully connected generator and discriminator networks for image synthesis tasks.
• Analyzed instability issues such as mode collapse and vanishing gradients during training.
• Visualized generated samples and monitored loss curves to evaluate convergence behavior.