Designed and developed this beautiful personal portfolio from scratch using Next.js and TypeScript. Optimized performance with server-side rendering for faster client response times and automated deployments via GitHub Actions and Vercel.

• Next.js
• TypeScript
• Vercel

Designed and built an embedded spatial measurement system using a ToF (Time-of-Flight) sensor to capture surrounding area data. Utilized Python NumPy and Open3D for 3D spatial mapping. Utilized UART and I2C for efficient software-hardware communication.

• Embedded C
• Python
• MCU
• I2C/UART

Collaborated in a team of three to design and develop an efficient Python program for the McMaster Recycling Plant Sorting System. Leveraged infrared (IR) technology to enhance sorting accuracy, improving it from 50% to 93%.

• Python
• Pandas

As a team of international students, we recognized the challenges newcomers face in Canada, from securing housing to opening a credit card. To address this, we developed a centralized platform that consolidates essential resources and information, simplifying the transition for incoming international students.

• React
• GoogleMapApi

Developed and integrated software and hardware modules for the McMaster AEV, a 1/10th-scale RC vehicle platform. Utilized Jetson Nano and ROS for the AEV’s operating environment, programmed the Electronic Speed Controller with the VESC tool, and implemented SLAM for simultaneous mapping and self-localization to enable autonomous driving.

• C/C++
• Python
• VESC

Designed and developed an advanced warehouse materials ordering system with real-time order display, detailed history tracking, and dynamic data synchronization. Optimized order processing, reducing cycle time by 70% and significantly enhancing operational efficiency.

• Next.js
• TypeScript
• Flask
• FastAPI

Developed a Smart Receiving Device for warehouse automation using C#, integrating the Keyence SR-5000X camera and multiple sensors to capture up to 8 barcodes per scan. Enhanced materials traceability and boosted production efficiency by 50%.

• C#
• Python

Optimized, profiled, and tested key machine learning operations (GEMM, GEMV, SPMM, SPMV) on the SciNet supercomputer using C++. Achieved a 5000x speedup by implementing cache-efficient algorithms and leveraging vectorization (AVX), CPU parallelism (OpenMP), and GPU parallelism (CUDA).

• C++
• Linux
• GoogeTest
• OpenMP
• CUDA

Developed a Lunar Lander agent using Deep Q-Networks (DQN) and Reinforcement Learning, achieving an average score of 265. Trained the agent in the OpenAI Gym environment with PyTorch for neural network implementation.

• Pytorch
• DQN/RL
• OpenAI Gym
• Neural Network

Developed an out-of-order CPU execution model with Reorder Buffer (ROB) and Load-Store Queue (LSQ) to optimize instruction processing. Designed instruction fetching, allocation, and commit mechanisms for parallelism. Improved performance with load forwarding to reduce memory latency. Analyzed IPC, ROB, and LSQ size to identify execution bottlenecks.

• C++
• Octopus(Cache Simulator)

Enhanced the xv6 operating system by implementing custom system calls and kernel-level features as part of MIT’s 6.S081 Operating Systems course. Developed user-space utilities (sleep, primes, find, xargs) and introduced debugging tools such as page table printing, backtracing, and lazy memory allocation.

• C
• Unix

Built an autonomous self-driving bike from scratch to address the last-mile bike sharing problem. Developed a Next.js web app with Google Maps integration, a FastAPI backend hosted on AWS EC2 using MQTT and Redis Pub/Sub for real-time communication, and PID-based motor control on an ESP32 with GPS data acquisition. Hardware integration included motor driver control, circuit design, and Bluetooth-based manual override. Awarded ECE department 3rd Place in the McMaster Capstone Competition.

• Raspi
• Full-Stack
• Linux
• Redis
• AWS EC2