Rahul Chauhan

I've built AI systems for food-image authentication, real-time gesture recognition, and Generative AI chatbots, taking models from data preprocessing and feature extraction through training, evaluation, and deployment.

At Cognifyz Technologies, I engineered end-to-end ML pipelines across 50,000+ restaurant records, training Random Forest and Linear Regression models that achieved 85%+ accuracy and reduced manual scoring effort by approximately 40%. I also improved Pandas/NumPy preprocessing, reduced data inconsistencies by 30%, optimized MySQL reporting time by 40%, and delivered 10+ stakeholder visualizations.

For Vision X, I built a CNN-based system to identify real versus fake or AI-generated food images, using visual feature extraction, texture analysis, and OpenCV image processing to support scalable content moderation for food-tech and marketplace platforms.

I've also built HandSpeak, a real-time gesture-recognition system using OpenCV, MediaPipe, and PyAutoGUI that classified gestures at sub-50ms latency and 30 FPS. This work earned first place at KIIT National Science & Technology Day and recognition in national AI and computer science competitions.