PersonaAI: Leveraging Retrieval-Augmented Generation and Personalized Context for AI-Driven Digital Avatars
Elvis Kimara, Kunle S. Oguntoye, Jian Sun · 2025
Read PersonaAI: Leveraging Retrieval-Augmented Generation and Personalized Context for AI-Driven Digital Avatars by Elvis Kimara, Kunle S. Oguntoye, Jian Sun (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-15489.txt.
Introduction
Elvis Kimara, Kunle Oguntoye, and Jian Sun introduce PersonaAI, a system that builds digital avatars imitating a real person's personality. Their paper came out in 2025. The system runs as a cloud-based mobile application.
Rather than training a huge model from scratch, the avatar draws on stored facts about the person at answer time, a method called retrieval-augmented generation. User data is collected through the app and kept in a secure database for that retrieval. A prompt-tuned open language model then answers in character, with context drawn from the person's own data. The authors argue this approach is lighter and more sustainable than full model training. They open-source their implementation to invite community development of personalized avatars.
Catalog record
| CREATOR | Elvis Kimara, Kunle S. Oguntoye, Jian Sun |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 36 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2503-15489
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
PersonaAI: Leveraging Retrieval-Augmented Generation and Personalized Context for AI-Driven Digital Avatars Elvis Kimara, Kunle S. Oguntoye, Jian Sun Abstract This paper introduces PersonaAI, a cutting-edge application that leverages Retrieval-Augmented Generation (RAG) and the LLAMA model to create highly personalized digital avatars capable of accurately mimicking individual personalities. Designed as a cloud-based mobile application, PersonaAI captures user data seamlessly, storing it in a secure database for retrieval and analysis. The result is a system that provides context-aware, accurate responses to user queries, enhancing the potential of AI-driven personalization. Why should you care? PersonaAI combines the scalability of RAG with the efficiency of prompt-engineered LLAMA3, offering a lightweight, sustainable alternative to traditional large language model (LLM) training methods. The system’s novel approach to data collection, utilizing real-time user interactions via a mobile app, ensures enhanced context relevance while maintaining user privacy. By open-sourcing our implementation, we aim to foster adaptability and community-driven development. PersonaAI demonstrates how AI can transform interactions by merging efficiency, scalability, and personalization, making it a significant step forward in the future of digital avatars and personalized AI. K eywords Retrieval-Augmented Generation \cdot Personalized AI \cdot Digital Avatars \cdot LLAMA \cdot Natural Language Processing 1 Introduction Artificial intelligence (AI) has become a transformative force across numerous domains, fundamentally reshaping the way humans interact with technology. While AI-powered virtual assistants such as Siri, Alexa, and ChatGPT have excelled in handling routine tasks and providing general-purpose assistance, these systems often lack the capacity to personalize interactions to reflect individual traits, preferences, and nuances. This inherent limitation has created a significant gap in user-centric AI solutions, where the interaction feels impersonal, static, and contextually inadequate. PersonaAI addresses this critical need by introducing a novel system capable of creating AI personas that mirror individual personalities. Designed to serve diverse user needs, PersonaAI offers applications ranging from preserving the digital legacy of loved ones to assisting professionals who require consistent and accurate information delivery in their absence. Unlike traditional large language models (LLMs), which rely heavily on users repeatedly providing context, PersonaAI employs Retrieval-Augmented Generation (RAG) to seamlessly capture, store, and retrieve user-specific data, enabling highly relevant, context-aware, and personalized responses . The need for personalization is especially urgent in today’s fast-paced and interconnected world. Generic AI interactions fail to engage users on a deeper level, diminishing user satisfaction and productivity. PersonaAI bridges this gap by utilizing a cloud-based mobile application that captures user data in real-time, securely storing it in a cloud database. By leveraging scalable and efficient LLMs such as LLAMA3 and combining them with advanced prompt engineering techniques , PersonaAI ensures a lightweight and cost-effective solution without the computational overhead of traditional LLM training. Beyond technical advancements, PersonaAI is built with a strong focus on user trust and reliability . Recognizing the challenges of AI hallucinations—generating responses that are incorrect or misleading—PersonaAI incorporates mechanisms to transparently acknowledge limitations. By explicitly responding with "I don’t know" when appropriate, the system mitigates inaccuracies and enhances user confidence. This approach is aligned with the ethical considerations necessary for deploying AI in sensitive and user-centric applications. This paper delves into the methodology, design, and implementation of PersonaAI, offering the following key contributions: 1. A Novel Approach to Personalization : PersonaAI combines RAG with LLAMA3 and advanced prompting to deliver tailored interactions that adapt dynamically to individual user profiles. 2. Scalable and Efficient Architecture : The system demonstrates the feasibility of using scalable LLMs with retrieval-based augmentation, significantly reducing the computational costs associated with training domain-specific AI models. 3. Innovative Data Collection through a Mobile App : A seamless and secure data ingestion mechanism captures user contexts while prioritizing privacy and usability. 4. Ethical Design and Reliability : PersonaAI enhances user trust through transparent handling of uncertainties, setting a benchmark for ethical AI design. By exploring these contributions, this paper aims to advance the understanding and practical deployment of personalized AI systems, highlighting their potential to transform digital interactions. The results and insights from PersonaAI lay the groundwork for broader applications in personalized education, healthcare, and digital legacy preservation. 2 Related Works Personalized AI systems have garnered significant attention in recent years, driven by advancements in natural language processing (NLP), retrieval-augmented generation, and memory-based neural architectures. Despite these advancements, existing solutions often fall short in delivering deeply personalized and scalable interactions, presenting opportunities for innovation. Large language models (LLMs) such as OpenAI’s GPT series [1] have redefined the capabilities of AI systems in generating coherent and contextually appropriate text. GPT-3 and GPT-4, for example, exhibit exceptional performance in natural language understanding and generation. However, these models typically operate as general-purpose assistants, requiring users to repeatedly provide context to produce personalized responses. This dependence on manual input limits their effectiveness in scenarios demanding dynamic and user-centric personalization. Retrieval-Augmented Generation (RAG), introduced by Facebook AI [2], addresses some of these limitations by combining external retrieval mechanisms with generative models. RAG enables the dynamic integration of context-specific information into the response generation process, significantly enhancing relevance and accuracy. Despite its promise, existing RAG implementations often focus on narrow, domain-specific applications, leaving a gap in its application to user-driven, personalized interactions at scale. Efforts to develop personalized virtual assistants, such as Google’s Meena [3] and Amazon Alexa [4], have incorporated user preferences and intent recognition to tailor interactions. These systems rely on static user profiles and predefined intent libraries, which can hinder their ability to adapt to nuanced or evolving user behaviors. Furthermore, their personalization strategies are typically limited to specific use cases, lacking the flexibility to generalize across diverse user contexts. Memory-augmented neural networks [5], such as those pioneered by Weston et al., have advanced the capability of AI systems to retain and recall historical user data. These networks facilitate long-term personalization by enabling systems to adapt to user preferences over time. However, challenges remain in scaling memory networks efficiently, particularly as the volume of stored user data grows, which can lead to computational bottlenecks and diminished response times. Ethical considerations have emerged as a critical area of research in personalized AI. Studies by Binns et al. [6] and Jobin et al. [7] emphasize the importance of transparency, data privacy, and user consent in AI systems that handle sensitive personal information. The deployment of personalized AI systems must prioritize secure data handling and provide users with control over how their data is captured and utilized. These principles are crucial for fostering trust and ensuring the responsible use of AI. Building on these foundational works, PersonaAI addresses key gaps in the field. By leveraging RAG in conjunction with LLAMA models, PersonaAI dynamically retrieves and utilizes user-specific contexts to generate accurate, personalized responses. Its innovative data collection approach, implemented through a cloud-based mobile application, ensures seamless and privacy-preserving context capture. Furthermore, PersonaAI explicitly addresses ethical concerns by embedding mechanisms to handle uncertainty and transparently acknowledge limitations, enhancing user trust and system reliability. In summary, PersonaAI builds on the progress of previous works while introducing novel methodologies to achieve scalable, secure, and deeply personalized interactions. Its contributions demonstrate the potential of AI systems to transform applications in education, healthcare, and beyond, paving the way for broader adoption of personalized AI technologies. 3 Experimental Platform This section details the infrastructure, algorithms, and deployment