PRANAV S SHETTY
Computer Science EngineerDeep LearningCybersecurity
B.E. in Computer Science Engineering at Sahyadri College of Engineering & Management. Specializing in the intersection of artificial intelligence and information security. Architecting resilient neural models and defending next-generation web infrastructure.
Transformer Lab
Tulu-to-English Machine Translation
Architected a custom NLP pipeline utilizing state-of-the-art Transformer self-attention mechanisms to translate the low-resource Tulu language. Trained on a curated corpus of 9,464 sentence pairs with sinusoidal positional encoding, achieving a 10-fold reduction in training computational power compared to LSTM/GRU.
Adversarial Vault
FiLM-Enhanced WGAN Password Gen
Developed a cutting-edge Conditional Wasserstein GAN augmented with Feature-wise Linear Modulation (FiLM) for synthetic password generation. Achieved superior length control and structural consistency compared to standard baselines like PassGAN.
Molecular Cluster
LogP Drug Lead Optimization
Engineered predictive models integrating Graph Neural Networks (GNN) and Random Forests to optimize the lipophilicity (LogP) of pharmaceutical drug leads. Benchmarked on 1,177 compounds from PubChem, identifying that the Random Forest algorithm (R²=0.84) outperformed the GNN approach.
The Medical Node
AI Agents for Cardiac Complications
A theoretical framework utilizing Autonomous AI Agents based on the ReAct paradigm to enable early detection of beta-blocker-induced cardiotoxicity. The conceptual architecture employs LLMs with Chain-of-Thought reasoning to transition clinical monitoring from passive alarms to active decision support.
Command Center
Web & Cybersecurity Intern
Conducted vulnerability assessments and security analysis for active infrastructure. Also architected a Full-stack College Stationary Management System featuring secure RBAC, dynamic JSP content rendering, and real-time MySQL inventory tracking.
Forensics Lab
Onyx Forensics Platform
A comprehensive digital forensics investigation platform built with TypeScript. Features advanced evidence analysis tools, forensic workflow management, real-time threat intelligence dashboards, and automated incident response pipelines.
MY TECH STACK
Interact with the physics cluster. Hover to scatter.