Project 01
Jul 2024 - Dec 2024
Real-Time Health and GPS Tracker for Climbers
This project started from a practical safety problem. Climbers in remote locations can
face sudden health risks, route changes, or communication issues, and I wanted to
build a system that could monitor critical signals and send alerts early.
- I combined health tracking and GPS monitoring into one safety-focused workflow.
- I used Raspberry Pi and Arduino UNO to handle sensing, monitoring, and device-level coordination.
- I worked on low-power thinking because remote use is not useful if the device drains quickly.
- I explored SVM-based analysis to better separate and study temperature and blood pressure patterns.
What mattered most in this project was not only collecting data, but making the system
meaningful in a real environment. That pushed me to think about reliability, alerts,
and how technical choices connect to user safety.
Project 02
May 2025 - Jul 2025
Music Genre Classification and Language Detection System
I built this project because it let me work at the intersection of machine learning,
audio analysis, and language understanding. I wanted to create something that did more
than classify a song, and instead combined genre recognition with lyric language
detection in one pipeline.
- I used Mel spectrogram features with Librosa to convert raw audio into useful model input.
- I built a CNN with TensorFlow and Keras for multi-class genre classification.
- I added OpenAI Whisper to transcribe audio and identify lyric language from the result.
- I used accuracy and loss tracking to evaluate model behavior and guide improvements.
This project helped me better understand how feature extraction, model selection, and
evaluation connect across an end-to-end ML workflow. It also pushed me to think about
how different AI tools can be combined into one useful system.
Other Project Writing I Want to Add
- Detailed case studies with screenshots, architecture notes, and final results.
- What went wrong during development and how I fixed those problems.
- How I chose tools, models, and workflows for different project types.
- Shorter blog posts about lessons from datasets, debugging, and iteration.
I want this page to keep growing into a stronger project journal, not just a static
summary page. As I build more work, I will add more focused write-ups here.