Pioneering AI Research for the Future
Explore our cutting-edge work in multimodal representations, agentic systems, language modeling, and computer vision
Our Research at Phronetic AI
We are a dynamic startup specializing in AI applications, combining expertise in traditional computer vision problems with state-of-the-art multimodal large language models (LLMs). Our innovative solutions address critical challenges in face recognition, object detection, tracking, and deep fake detection. Additionally, we pioneer agentic flows, creating intelligent assistants capable of performing various managerial roles.
Multimodal Representations
We aim to build AI models capable of conversing with multimodal content, with focus on efficiency and performance. We have already announced Owlet, a family of models for video question answering, which process both visual as well as audio signals. Our research focuses on model architectures, as well as datasets spanning multiple modalities. We are also conducting research on interesting problems like video search, video-language grounding, video embeddings, activity detection, etc. leading to an expertise in video understanding in general.
Agentic Systems
We focus on creating intelligent agents that excel in a variety of tasks, including task management, observational learning, and visual planning. Our task manager agents are designed to efficiently organize, prioritize, and execute tasks, enhancing productivity and workflow automation. Through our "watch and learn" agents, we enable systems to observe human actions and learn from them, improving their ability to perform complex tasks autonomously. By integrating these capabilities into our platform, we aim to build versatile, intelligent agents that can seamlessly assist in diverse real-world applications, driving innovation and efficiency.
Language Modeling
Our team specializes in fine-tuning large language models (LLMs) for custom use cases, with a strong emphasis on multilingual datasets and code generation. We have also fine-tuned LLMs specifically for code generation tasks, optimizing them to produce high-quality code across various programming languages. Through our innovative approaches, we aim to expand the practical applications of LLMs, making them more adaptable and effective for specialized industry needs.
Computer Vision
Our team is dedicated to advancing the field of computer vision, with a focus on solving critical problems such as face recognition, face re-identification (re-id), object detection, object tracking, and the detection of deep fakes. Our research aims to develop cutting-edge technologies that can be applied across various industries, enhancing security, efficiency, and user experience.
Research Blogs
Join Us in Shaping the Future of AI
Collaborate with our innovative research team or apply to help drive groundbreaking advancements in AI technology
