Background
Revealed
👋 Welcome to my portfolio

Shivam Kashyap

• AI-ML Enthusiast •

Crafting intelligent systems with deep learning and computer vision. BitBox 5.0 Hackathon Winner. Specializing in YOLOv8, PyTorch, and edge AI deployment.

About Me

I am a Electronics and Communication student at Jaypee Institute of Information Technology with a passion for building intelligent systems that solve real-world problems. My journey in machine learning and computer vision has led me to develop cutting-edge solutions in object detection, image classification, and AI deployment.

My expertise spans from training deep neural networks with PyTorch and TensorFlow to deploying edge AI models on resource-constrained devices like Raspberry Pi. I combine strong theoretical understanding with practical implementation skills to create production-ready AI systems.

Recently, I won 1st place at BitBox 5.0 Hackathon for SahYatri, an IoT-enabled real-time bus occupancy monitoring system using YOLOv8 object detection. This victory validated my ability to build scalable, impactful AI solutions.

3+
Major ML Projects
BitBox
Hackathon Win
8.5/10
University CGPA
6+
Tech Stacks

Technical Skills

Programming Languages

PythonC/C++JavaScriptTypeScriptSQLPHP

ML/AI Frameworks

PyTorchTensorFlowKerasYOLOv8ResNetMobileNetOpenCV

Web Technologies

React.jsNode.jsNext.jsFlaskStreamlitTailwind CSS

Databases & Cloud

MongoDBPostgreSQLMySQLAWSDockerGit/GitHub

Specializations

Computer VisionDeep LearningObject DetectionImage ClassificationNLPTransfer Learning

Hardware & IoT

Raspberry PiArduinoCamera ModulesSensors

Featured Projects

SAHYATRI — IoT Bus Occupancy Monitoring

🏆 BitBox 5.0 Hackathon Winner

Real-time public transport monitoring system using YOLOv8 object detection deployed on Raspberry Pi for edge inference.

→YOLOv8 model achieving 95% accuracy on 1000+ test images for passenger counting
→Real-time inference at 15-20 FPS with <100ms latency on Raspberry Pi 4
→Multithreading pipeline for parallel video processing and LCD display updates via GPIO
YOLOv8PyTorchIoTRaspberry PiPython
GitHub

Food Recognition & Nutritional Analysis AI

Deep learning system for food recognition and comprehensive nutritional analysis with API integration.

→ResNet-50 CNN trained on Food-101 dataset (101,000 images) with 92% top-1 accuracy
→Data augmentation pipeline increasing dataset by 300% with 15% overfitting reduction
→Gradio interface deployed on Hugging Face Spaces with macro/micronutrient breakdown
ResNet-50TensorFlowGradioAPI Integration
Demo

Mathematical Expression Solver

Handwritten equation recognition and algebraic solving system with OCR and symbolic computation.

→Fine-tuned EasyOCR model achieving 89% recognition accuracy across 14 character classes
→OpenCV preprocessing pipeline with 85% character segmentation accuracy
→Flask REST API integrating SymPy and NumPy for expression parsing and solving
EasyOCROpenCVFlaskSymPyPython
GitHub

Ready to collaborate?

I'm always interested in discussing new projects and opportunities. Let's connect and create something amazing together.