Gowtham
Profile
Passionate and detail-oriented ML Engineer and Vibe Coder with a strong background in designing, developing, and optimizing intelligent systems. Currently pursuing B.Tech in Information Technology at VIT University. Experienced in building predictive models, deep learning frameworks, and end-to-end machine learning pipelines. Specialized in computer vision, time-series forecasting, and creating real-world ML applications that solve practical problems. Strong foundation in both classical machine learning algorithms and modern deep learning architectures.
Education
Bachelor of Technology - Information Technology
VIT University, Vellore
Relevant Coursework: Machine Learning, Deep Learning, Data Structures, Algorithms, Database Management Systems, Web Technologies, Software Engineering
Pre-University College (PUC)
Narayana E-Techno School
Secondary School (Grade 10)
BMD Jain School
Projects
Predictive Maintenance Framework for Metro Train Systems
Developed a comprehensive predictive maintenance framework leveraging classical regression models and hybrid deep learning architectures for accurate failure prediction and remaining useful life (RUL) estimation in metro train Air Production Units. The system analyzes real-world sensor data to predict equipment failures before they occur, reducing downtime and maintenance costs.
Key Features:
- Implemented hybrid deep learning models combining CNN and LSTM for sequential pattern recognition
- Achieved 92% accuracy in failure prediction using multi-sensor fusion
- Developed RUL estimation algorithm for proactive maintenance scheduling
- Real-time anomaly detection system for immediate fault identification
- Data preprocessing pipeline handling missing values and sensor noise
Banana Ripeness Prediction System
Created an intelligent computer vision system for real-time banana ripeness classification designed for real-life deployment in agricultural and retail environments. The system uses deep learning to classify bananas into different ripeness stages, helping optimize harvest timing and reduce food waste.
Key Features:
- Custom CNN architecture trained on 5000+ banana images
- Multi-class classification: Unripe, Ripe, Overripe stages
- Real-time inference with 89% accuracy
- Deployed on edge devices for on-site usage
- Integrated with mobile application for farmer accessibility
Automated Waste Classification System
Developed an AI-powered waste classification system for real-world deployment in waste management facilities. The system automatically categorizes waste into recyclable, organic, and non-recyclable categories using computer vision, promoting sustainable waste management practices.
Key Features:
- Transfer learning using pre-trained ResNet50 model
- Three-class classification: Recyclable, Organic, Non-recyclable
- Data augmentation techniques for robust performance
- Achieved 87% classification accuracy across diverse waste types
- Optimized for deployment on low-power devices
- Integrated with conveyor belt systems for automated sorting
UI/UX Design for Microworms.in
Designed and implemented a modern, responsive user interface for Microworms.in, an e-commerce platform specializing in aquarium supplies. Created an intuitive shopping experience with focus on visual appeal, user engagement, and conversion optimization.
Key Features:
- Modern, clean interface with intuitive navigation
- Mobile-first responsive design for all screen sizes
- Product showcase with high-quality image galleries
- Smooth animations and transitions for enhanced user experience
- Optimized loading times and performance
- User-friendly cart and checkout process
- Integrated product filtering and search functionality