Parth Gupta

Developing an AI-based dermatology application for 15 skin disease conditions, utilizing image classification techniques with synthetic data, transfer learning, and vision transformers to accurately diagnose health conditions, achieving overall 95% accuracy rate. Designed and implemented a multimodal model combining image and text data, applying unsupervised clustering algorithms for feature extraction and text embedding, alongside supervised few-shot classification and deep Siamese networks resulting in a 40% improvement in model performance. Leading successful integration of LLM technologies (RAG, Lang chain, Pinecone, transformers) to query and analyze data from 10 websites, demonstrating advanced data retrieval and processing capabilities. Engineered a robust Model API on Google Cloud for an AI-based diagnosis app, significantly improving diagnostic accuracy. Developing and deploying scalable backend web applications using Django REST Framework, 50 % operational efficiency and enhancing user experience. Mentoring and leading 5 AI interns during project execution by providing constructive feedback on model training to enhance team productivity and innovation. Facilitated the acquisition of INR 10 crores in ICMR grants by guiding medical professionals in AI research proposals and tutored over 100 B.Tech students in Python, blending practical AI applications with theoretical concepts.

Engineered a highly accurate model utilizing Natural Language Processing (NLP), Transformers, and Artificial Neural Networks (ANN) to classify textual reviews of vegan and non-vegan foods. Extracted discriminant words for each class,enhancing the interpretability of the model. Applied advanced techniques including tokenization, lemmatization, named-entity recognition (NER), part-of-speech (POS) tagging, and innovative feature engineering to pre-process reviews, ensuring robust performance.

Spearheaded the development of a predictive model leveraging ATS Data, achieving an impressive 94.6% accuracy after meticulous tuning. This model revolutionized visibility predictions, improving strategic decision-making. Strategically employed machine learning and data analytics on the Uber dataset for market segmentation, driving insights into customer behavior and preferences.