Skip to content
View gurshh-rain's full-sized avatar

Achievements

Achievement: YOLO

Achievements

Achievement: YOLO

Highlights

  • Pro

Block or report gurshh-rain

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
gurshh-rain/README.md

Mechatronics Engineering student at the University of Waterloo (Tron '31). Building with passion in robotics, computer vision, and AI.
I like making things that actually move, think, and work in the real world.

gurshh-rain's GitHub profile

🤖 What I Build

Hardware & Robotics

  • Bionic Hand — 3D-printed prosthetic hand controlled by real-time hand tracking via MediaPipe, mapping joint angles to servo positions for fluid, natural motion.
  • Quadpod Robot — Four-legged walking robot with PID-controlled coordination across all legs simultaneously. Used ultrasonic sensor for object detection while walking.

Controls & Systems

  • Automatic PID Tuner — Uses gradient descent to automatically tune PID controller gains in MATLAB/Simulink, eliminating manual trial-and-error tuning.

Computer Vision & ML

  • Turbofan RUL Prediction — Predicts Remaining Useful Life of jet engines using the NASA C-MAPSS dataset, comparing Random Forest, XGBoost, LSTM, and an LSTM+XGBoost ensemble.
  • iEEG Signal Transformer (ongoing) — Transformer architecture for predicting iEEG data from neural signals.
  • Framelyai — Behavioral Interview Analyzer — Web platform analyzing eye contact, posture, expressions, and filler words in real time, with an LLM grading answers against the STAR framework.
  • Gesture-Controlled Computer — Real-time hand tracking to click, scroll, and drag using only gestures, built with MediaPipe + PyAutoGUI.
  • Satellite Deforestation Detector — Custom CNN classifying satellite imagery to detect deforestation, trained on the Amazon from Space dataset.

AI & Agents

  • RAG Portfolio Chatbot — Ask anything about me, built from scratch with sentence-transformers, FAISS vector search, and a local LLM.
  • Linked Zip Bot — Reinforcement learning agent trained to solve the Zip game.

🛠 Skills

Languages  Python C++ MATLAB Java

ML / AI  PyTorch scikit-learn OpenCV MediaPipe Pandas NumPy Matplotlib

Robotics  ROS2 PID Control Kinematics Servo/Motor Control

📫 Let's Connect

Always building something. Always learning.

Pinned Loading

  1. Modelling-LSTM-Random-Forest-and-XGBoost-to-Forecast-RUL-Metrics-of-NASA-Turbofan-Jet-Engines Modelling-LSTM-Random-Forest-and-XGBoost-to-Forecast-RUL-Metrics-of-NASA-Turbofan-Jet-Engines Public

    Predictive maintenance for turbofan jet engines: estimates Remaining Useful Life (RUL) using Random Forest, XGBoost, and LSTM, plus an optimized ensemble. Built on NASA's C-MAPSS dataset with rolli…

    Jupyter Notebook

  2. Time-Series-Forecasting-of-Household-Power-Usage Time-Series-Forecasting-of-Household-Power-Usage Public

    Forecasts household energy consumption (kW) using time-series feature engineering (lags, calendar features) and a hyperparameter-tuned XGBoost regressor. Trained on the UCI Individual Household Ele…

    Jupyter Notebook

  3. Multi-Temporal-Land-Cover-Classification-and-Deforestation-Detection-via-Convolutional-Neural-Networ Multi-Temporal-Land-Cover-Classification-and-Deforestation-Detection-via-Convolutional-Neural-Networ Public

    Trains a CNN on EuroSAT satellite imagery for land-cover classification (93.6% test accuracy), then applies it to real GeoTIFF scenes across two time points to automatically flag Forest → non-fores…

    Jupyter Notebook

  4. Gradient-Based-Auto-Tuning-of-PID-Controller-Gains-via-Numerical-Optimization Gradient-Based-Auto-Tuning-of-PID-Controller-Gains-via-Numerical-Optimization Public

    Implements a PID controller from scratch and auto-tunes its P/I/D gains using gradient descent with numerically estimated gradients, minimizing a loss combining tracking error and overshoot.

    Python

  5. ML-Algorithms-Python-Implementation ML-Algorithms-Python-Implementation Public

    Machine learning algorithms implemented from scratch in Python to build a deeper understanding of their mathematical foundations and inner workings. Includes clear, well-documented implementations,…

    Python

  6. gesture-controlled-computer gesture-controlled-computer Public

    A Python-based computer vision application that uses MediaPipe, OpenCV, autopy, and pyautogui to allow real-time control of your PC's mouse cursor, clicking, dragging, and adaptive scrolling via ha…

    Python