π€ Making AI so smart, it might take my job someday.
πΉ Passionate AI/ML Developer | LLM Enthusiast | Generative AI | Deep Learning
πΉ Exploring the cutting edge of AI with Large Language Models & Computer Vision
##π§ Currently Working On:
-
π **Large Language Models (LLMs): Fine-tuning GPT-2 for text generation, exploring transformer-based architectures.
-
π― **Image Recognition & Face Recognition: Built EfficientNet models (B0, B4, B7) for a face recognition attendance system.
-
π **Image Steganography: AI-powered techniques for secure data hiding & privacy-preserving communication.
-
π **Kaggle Competitions: Participating in Playground Series S4E11 (mental health analysis).
-
π** Deep Learning Architectures: Exploring advanced architectures like EfficientNet, Inception-ResNet-v2, and Vision Transformers to improve accuracy and performance in image classification tasks.
-
π― **Computer Vision & Object Detection: Implemented object detection models such as YOLOv5 and EfficientDet for real-time applications, including assistive technologies and smart surveillance.
-
π€ **Generative AI & NLP: Gaining hands-on experience in fine-tuning GPT-2 and building custom transformer models for conversational AI, chatbots, and content generation tools.
-
π‘ **Cybersecurity & Malware Detection: Designing a deep learning-based malware detection system focused on classifying Android malware, especially Trojans, using behavior-based and static features.
-
β‘**PyTorch & TensorFlow (tf_keras): Advancing model training using custom loops, quantization, and deployment optimization.
-
π§ **LLMs & Prompt Engineering: Exploring instruction tuning, few-shot learning, and real-world LLM applications.
-
π **Advanced NLP Techniques: Training transformers for text summarization, question answering, and dialogue systems.
-β **Developed an AI-powered face recognition attendance system with real-time processing using EfficientNet. -β **Built AI-enabled smart glasses with IoT, object detection, and real-time voice alerts for the visually impaired. -β **Designed an NLP-based career assistance chatbot with intelligent query handling and mentor/job matching features. -β **Achieved 92% accuracy with 40% faster inference in a facial emotion recognition system. -β **Created a fake news detection model using NLP with 90% accuracy and real-time classification. -β **Published research on AI-driven image steganography for secure data transmission and privacy protection. -β **Coordinated university-level hackathons and tech innovation events as part of Team Tech Nova at Sharda University.