Distributed E-commerce Inventory System (Database & Backend)
10/01/25, Windsor, ON
Designed and implemented a distributed e-commerce inventory backend using MySQL, Redis, Docker, and Kubernetes, supporting multi-node replication and fault tolerance.
Built ACID-compliant purchase workflows with row-level locking to prevent overselling under high concurrency.
Created Prometheus + Grafana dashboards to monitor query latency, lock waits, and replication lag; achieved P95 latency under 90 ms during load tests.
Developed Python-based multi-threaded client simulator generating 100+ concurrent user sessions for performance benchmarking.
Semi-Supervised Learning for Imbalanced Tabular Data (Machine Learning Project)
11/01/25, Windsor, ON
Implemented OS-level process management utilities in C using fork(), execvp(), pipe(), pthreads, signals, and file descriptors.
Built a mini shell supporting piped commands, background jobs, signal handling, and parent/child process synchronization.
Developed a backup automation system using tar, differential backup logic, timestamp tracking, and custom logging, achieving reliable recovery across multiple states.
Debugged race conditions and deadlocks; applied 2-phase locking and precedence graph reasoning to ensure schedule serializability.
