Hi, I'm Vahag!


Software Engineering Student at SJTU

I am a forth-year Bachelor's student at Shanghai Jiao-Tong University. My research interests lie in Cloud and Mobile Computing, Operating Systems, and MLSys. Besides tech, I have a huge passion for stand-up comedy.

Research

smart-madvise

Smart-madvise is a kernel module which collects real-time memory usage data, allowing the kernel to make more informed decisions for memory optimization across multiple applications, rather than relying on static hints provided by individual processes.

eBPF-Based Information Flow Control

eBPF-based security enhancement system built on a forked version of Bouheki, extended to dynamically monitor and enforce information flow policies via IPC channels, such as signals and sockets.

Projects

minik8s

Lightweight Kubernetes-like container orchestration tool, built from scratch in Go, featuring full CI/CD pipeline integration and comprehensive testing.

Online Shop Platform

Developed a comprehensive online shopping platform with extensive tooling and frameworks. Implemented React.js for the front-end and Spring Boot for the back-end. Managed structured and unstructured data using MySQL and MongoDB. Utilized Kafka for asynchronous messaging and Redis for effcient caching. Employed Neo4j for graph-like data structures, Apache Spark for keyword count tasks, Nginx for load balancing and Docker for containerization

cloudstoragio

Created a Python library that unifies file operations for AWS S4, Google Cloud Storage, and Dropbox, mimicking local file system commands for cloud storage management.

Distributed File System

Engineered an inode-based, distributed file system in C, while emphasizing concurrency, atomicity, and crash recovery features for robustness.

Raft and Map Reduce

Implemented the Raft consensus algorithm and the MapReduce programming model in C by following their respective academic papers.

Tiger-Compiler

Developed a complete compiler for the Tiger programming language in C++, covering all key stages from lexical analysis to register allocation.

Video Classification

Created a Video Classification System using Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) networks.

Event Management Platform

In a collaborative class project, led the development of a school event management website, guiding the project from inception to completion while upholding fundamental software engineering and project management principles.

Computer System Labs

Completed a series of complex labs, where I reverse-engineered a binary program, simulated attacks using buffer overflow bugs, build a simulator and assembler for the Y86-64 instruction set, developed a simple web proxy and Unix shell.

LeetCode Solutions

Successfully tackled over 50 medium to hard challenges on LeetCode using C++, complemented with thorough explanations, while maintaining clean and efficient code.

Certifications

Machine Learning Specialization - DeepLearning.AI

Deep Learning Specialization - DeepLearning.AI

TensorFlow Developer - DeepLearning.AI

Fundamentals of Accelerated Data Science - Nvidia