AI Engineer building Scalable LLM & ML Systems | Generative AI • MLOps • Real-Time Data | AWS

Vijaykumar Shivakumar Kalasa  • Peerlist

• Developed a vision–language pipeline using BLIP-2 and LLaVA to detect accessibility hazards from images, applying spatial reasoning and prompt conditioning to generate clear, context-aware safety explanations for visually impaired users, evaluated through human feedback

Adaptive Vision–Language Reasoning forAccessi

• Built a hybrid SSDAE and CNN–LSTM model to detect DDoS attacks in SDN by analyzing network traffic, extracting spatial and temporal patterns, and achieving approximately seventy-nine percent accuracy on benchmark datasets through effective feature learning techniques.