About Me

Schulich Leader Scholar at McMaster University, one of Canada's top 50 entrepreneurial engineers that was awarded $120,000.
Engineered a biologically inspired recurrent vision model that outperformed MIT’s CORNet-S in adversarial robustness across MNIST, CIFAR, and ImageNet100.
Worked with Synaptive Medical and real surgeons to develop a surgical-vision system trained on live operating-room data, achieving 95% tool segmentation accuracy for real-time surgical guidance.
Founded Hack49 Global, a worldwide hackathon with 950+ participants from 40+ countries, securing $19,000+ in sponsorships
Built Sentinel, a hardware-aware static analysis system with a custom language server, performing RAM and Flash estimation, pin-mapping, and 50+ embedded validations before compilation.

Technical Skills

Programming Languages

Python, JavaScript, TypeScript, Java, C#, C++

Machine Learning

PyTorch, Keras, MATLAB

Web Development

React, Flask, TailwindCSS, Electron

Computer Vision

OpenCV, ITK SNAP

Databases

SQLite

Embedded Systems

Arduino, Raspberry Pi, ESP32

Tools & Platforms

Unity, Git/GitHub, VS Code

3D & Design

Fusion 360, Autodesk Inventor, Blender, SketchUp, OnShape, KiCad, CURA

Featured Projects

A showcase of my latest work in machine learning, computer vision, and robotics.

From Pixels to Precision

Developed a neural network that segments surgical tools from minimally invasive surgical videos, improving surgical precision and safety.

Adversarial Robustness Research

Developed a novel ventral visual neural network, based on CORnet-S architecture with denoising and gated recurrent blocks, achieving robustness of 97.87% under evaluated model threats.

Sentinel

The first hardware-intelligent static analyzer for embedded systems. Predicts RAM, pin, and protocol issues instantly via a custom VS Code + LSP toolchain.

Adaptive Learning Robot

A dual-system robot inspired by Daniel Kahneman's System 1 / System 2 model that learns new skills instantly from natural language, without pre-training or rigid scripts. Built with Cohere LLMs, real-time search, and 2 custom 3-DOF robotic arms, the system transforms messy instructions into smooth, executable motions. Placed Top 32 out of 256 projects at Hack the North, Canada's Largest Hackathon.

Zephyr AI Assistant

Built a keyboard-activated AI assistant that seamlessly integrates into your workflow with a single hotkey press. Leverages Groq's Llama 3.1 8B model to understand natural language commands and instantly control Spotify playback, check weather conditions, aggregate news feeds, manage calendar events, and track project tasks.

DermAI Medical Assistant

Built a DenseNet121 model trained on 19,000+ DermNet images with integrated RAG chatbot for dermatological triage assessment.

Pose Estimation Model

Implemented a Keypoint R-CNN model with OpenCV for real-time pose estimation in security footage and surveillance applications.

Gesture Controlled Robotic Arm

Developed a robotic arm controlled by hand gestures using computer vision and Arduino, demonstrating real-time human-robot interaction.

View More Projects

Rishi Shah

Research Engineer focused on neural networks and their real-world applicability.