RISHI JAIN

Led research into Agentic AI paradigms, focusing on enhancing GPT-style large language models through In-Context Learning and co-training methods.

• Investigated Agentic AI paradigms, focusing on improving the robustness and adaptability of GPT-style large language models (LLMs) using In-Context Learning (ICL) and co-training methods.
• Developed and implemented non-parametric enhancement strategies that boosted LLM performance without gradient-based fine-tuning.
• Leveraged contextual prompts and multi-source knowledge transfer to optimize LLM adaptability and robustness.

Engineered and deployed advanced machine learning solutions, enhancing OCR accuracy and developing a Retrieval-Augmented Generation system.

• Fine-tuned the SOTA PARSeq OCR model for CAPTCHA recognition using a custom curated dataset of over 2 million samples, boosting accuracy from 52% to 88%.
• Deployed the enhanced OCR model into production using TorchServe and Docker, ensuring scalable and efficient performance.
• Developed a Retrieval-Augmented Generation (RAG) system, including an effective evaluation technique to measure and optimize its performance.

Applied advanced computer vision and LSPIV techniques to analyze drone footage for environmental monitoring and developed a CNN model for velocity estimation.

• Applied computer vision and Large Scale Particle Image Velocimetry (LSPIV) techniques to analyze drone footage for flood and river velocity measurement.
• Developed and trained a Convolutional Neural Network (CNN) model in PyTorch for precise velocity estimation.
• Utilized extracted frames and advanced Computer Vision (CV) methods to enhance the accuracy of environmental data analysis.

Designed and implemented a novel model-agnostic method for time series classification, significantly enhancing ML model interpretability.

• Designed and implemented TS-NUC, a novel model-agnostic method for counterfactual explanation in time series classification.
• Achieved superior performance on UCR datasets by leveraging LSTM-Autoencoders and latent space optimization.
• Enhanced transparency in ML models by generating actionable counterfactuals for both multivariate and univariate time series, bridging the interpretability gap in black-box classifiers.

Developed an AI-powered platform for vocational education and career development, leveraging GenAI and innovative VR controls for enhanced user experience.

Engineered a medical imaging project for precise tumor detection and segmentation using deep learning, integrated with real-time Augmented Reality visualization for surgical planning.

Designed and implemented an AI-powered financial forecasting platform using LSTM models, real-time sentiment analysis, and fuzzy logic for enhanced market prediction accuracy.

Awarded for 'Adding Relevant Vocational Training into Elementary and Secondary Education Curriculum', demonstrating innovation in educational technology.

Recognized for 'Heritage Odyssey: Improving Attraction through Immersive Technology', showcasing creativity and technical skill in immersive experiences.

Published research on an AI-powered immersive virtual reality laboratory designed for pedagogical applications.

Published research on a novel model-agnostic method for counterfactual explanation in time series classification, enhancing ML model interpretability.