Jayashree Karlekar
Papers
1
Total Citations
5
H-Index
1
About
Dr. Jayashree Karlekar is a leading researcher in multimodal AI and embodied agents, with a focus on bridging large language models (LLMs) with real-world decision-making. Her most notable contribution is the development of **RAP (Retrieval-Augmented Planning with Contextual Memory)**, a groundbreaking framework that enables multimodal LLM agents to leverage past experiences for more intelligent, adaptive planning. By integrating a contextual memory system with retrieval-augmented generation, RAP allows agents to reflect on prior actions—mimicking innate human behavior—to improve performance in complex domains like robotics, gaming, and API integration. This work, published in 2024 and already garnering 5 citations, addresses a critical gap in AI: the ability to learn from history without catastrophic forgetting. Dr. Karlekar’s research directly tackles the challenge of making AI agents not just reactive, but truly reflective and efficient in dynamic environments. Her innovative approach to memory-augmented planning positions her at the forefront of next-generation autonomous systems, with implications for everything from household robots to enterprise automation.
Research Focus
Key Achievements
Top Papers
- 1