Projects from coursework, internships and those done in my own time
Built dual neural network implementations analyzing 70,000+ fashion product images using NumPy and PyTorch-Torchvision. Achieved 85%+ accuracy in classifying clothing items through custom backpropagation and optimization techniques.
Built a recommendation system analyzing 1M+ book ratings using collaborative filtering and K-means clustering in Python. Implemented user-based similarity metrics and cluster analysis for personalized book suggestions.
Engineered a medical diagnostic model analyzing patient health metrics using NumPy-based logistic regression and perceptron algorithms. Implemented gradient descent and MLE for 90%+ prediction accuracy on clinical data.
Made my own terminal that handles bash and unix commands, including wildcards, subshell, scripts, etc. Utilized lex and yacc for parsing, and implemented the terminal using C and C++ with extensive use of system calls.
Analyzed Fitbit time-series data using pandas and seaborn, performing statistical analysis on health metrics to discover activity patterns through data visualization and correlation studies.
Implemented TF-IDF vectorization and NLTK-based NLP pipeline in Python to analyze document similarities across classic literature, featuring custom preprocessing and cosine similarity calculations.
Built a C compiler translating to x86_64 assembly, utilizing lex and yacc for robust parsing, supporting diverse C features.