Rahul Bhoyar

Led cutting-edge AI research projects including DriveAgent, an LLM-driven multi-agent autonomous driving system combining sensor fusion and reasoning models for enhanced decision quality. Developed multi-agent scientific literature assistant TPRS leveraging retrieval-augmented generation and graph traversal techniques. Focused on using LangChain, LangGraph, HuggingFace models, and cloud deployments on AWS with containerized microservices. Implemented end-to-end pipelines for real-time inference, multi-modal sensor integration, and agent orchestration with observability and self-correction loops.

Developed a generative AI system for user acquisition and gamification with synthetic user behavior generation using RAG pipelines and fine-tuned LLM chatbots for engagement. Implemented fraud detection algorithms using GANs and VAE embeddings. Built scalable data pipelines and deployed microservices on AWS leveraging Sagemaker and Bedrock. Optimized recommendation and fraud detection models with metrics such as Precision, AUC-ROC, and F1 scores.

Implemented advanced textual and semantic analysis frameworks for news portals and financial portfolio optimization. Built robust data pipelines using Python, PySpark, and AWS services. Developed machine learning recommendation systems incorporating sentiment analysis and real-time data to suggest personalized investment strategies. Used technology stacks including TensorFlow, Pandas, Scikit-learn, and AWS Sagemaker for scalable deployment.