Shift.

A high-performance LSM-tree–based storage engine written in Java. Designed to explore compaction strategies, sparse indexing, and efficient on-disk layouts for modern workloads.

Current Progress

• Memtable and WAL integrated (by collaborator)
• Implementing SSTable writer, sparse index, and metadata tracker
• Compaction manager for leveled strategy
• Benchmarking and visualization of segment metadata
• Adaptive compaction experimentation

A lightweight machine learning project where I ported an existing logistic regression pipeline to a Naive Bayes–based classifier. Focused on comparing model performance and a usable outcome, to integrate with Nycto OS.

Current Progress

• Completed port from logistic regression to Naive Bayes
• Evaluating model performance

Technical Challenges

Balancing model accuracy and runtime performance while re-implementing the pipeline from scratch and documenting findings .