Work in Progress
Research-driven systems, relentless iteration, and intelligent experiments evolving into real-world AI solutions.
After nearly a year of research, experimentation, and iterative refinement, I architected an explainable ensemble deep learning framework for brain tumor detection using ResNet50, EfficientNet-B0, and DenseNet121 integrated with Grad-CAM. The challenge involved balancing model diversity with interpretability while ensuring consistent clinical relevance across architectures. Competing against 60 national teams, this experience strengthened my research depth, analytical rigor, and ability to communicate complex AI systems with clarity and confidence.
Selected among 100 colleges nationwide from over 10,000 delegates, I represented MIT Mysore as the only institution from Mysuru district to qualify. The evaluation process required technical clarity, feasibility validation, and strong articulation of innovation impact. Translating AI-driven solutions into a compelling, time-bound pitch under competitive scrutiny sharpened my ability to bridge engineering depth with strategic thinking and real-world applicability.
Won a 24-hour national hackathon against 120+ teams by engineering and optimizing a breast cancer detection model using thermal and ultrasound imaging. Under intense time constraints, the key challenge was ensuring strong generalization while handling noisy medical data and avoiding overfitting. This experience reinforced rapid decision-making, structured experimentation, and the ability to deliver reliable AI systems under pressure.
DL Intern at
Runshaw.in
Developed strong proficiency in designing, training, and optimizing deep neural networks for real-world image analysis and predictive tasks, translating theoretical concepts into production-oriented solutions.
Strengthened my understanding of data preprocessing and pipeline structuring, learning to handle noisy, imbalanced, and imperfect datasets that significantly impact model performance.
Gained hands-on experience in model evaluation and fine-tuning, addressing challenges like overfitting, convergence instability, and performance trade-offs through systematic experimentation.
Improved my ability to debug, iterate, and optimize under practical constraints, building resilience and disciplined problem-solving skills essential for industry-grade AI development.