Technology – New Wand

Wand AI has a dedicated team of industry-leading scientists focused on advancing AI reasoning and personalization. Our innovative technologies enable us to create more intuitive and customized AI solutions to drive real business impact.

View Open Roles

Our Vision for the Future

At Wand Research, we are revolutionizing human-machine collaboration for enterprise applications by building personalized and intelligent systems.

We are enabling a future where billions of AI assistants will manage knowledge-based tasks, freeing humans to focus on strategy and innovation. This would dramatically boost GDP and completely transform the software industry.

Focus Areas

We are doing pioneering research in the areas of reasoning and personalization — with the objective of enhancing AI intelligence and adaptability.

Reasoning & Wisdom Team

This team imbues AI systems with wisdom, mimicking human cognition. We develop advanced algorithms and models that enable AI agents to perform complex reasoning, understand context, and make intelligent decisions by considering future outcomes.

We are developing modular, agent-based architectures that enable advanced thinking, planning, and reasoning beyond current prompt engineering or model fine-tuning.

The interconnected agents guide core AI systems to achieve desired outcomes while avoiding undesirable behaviors, with the ability to adapt to changing goals in real-time.

Personalization & Adaptation Team

This team is developing AI systems that truly understand users. We adapt AI behavior to individual preferences, ensuring uniquely relevant support for each user, aiming for intuitive, human-like collaboration.

We integrate history analysis, real-time adaptation, and continuous learning for personalized interactions. Our models use comprehensive profiles and custom datasets for context-aware responses.

We build evolving AI systems for more effective user engagements, with the aim of revolutionizing AI personalization.

Research Leads

Our research teams are led by industry thought leaders with deep expertise in developing powerful, business-focused AI technologies.

Dr. Mehdi Fatemi

Team Lead: Reasoning & Wisdom

Dr. Mehdi Fatemi is a research lead, senior AI researcher, and innovator. He leads the Reasoning & Wisdom Team at Wand Research. Prior to Wand, he was one of the founding members of the Reinforcement Learning (RL) Team at Microsoft Research Montréal.

Dr. Fatemi’s goal is to add *wisdom* and *insight* to AI systems and his work centers around exploring the pivotal role of "value" in sequential decision-making and reasoning. Over the past 14 years, he has played key roles in numerous successful projects, which include: invention of dead-end analysis and process- based causality; state-of-the-art detoxification of Large Language Models through far-sighted reasoning; value mapping in RL; separation of concerns (SoC) in RL; world-record in Ms. Pac Man; RL for dialogue systems; and cognitive control.

Dr. Fatemi’s research has been featured in over 100 news outlets including Time , BBC, Fortune Magazine, MIT’s Homepage, MIT News, Cover-story of Proc. IEEE, Microsoft Research Blog, Business Insider , Forbes, The Verge, and many more. He is a senior member of the IEEE and has been named as an inventor on five issued/filed US patents in RL subjects. He has served as area chair, reviewer, and meta reviewer at all the top AI venues including NeurIPS, ICML, ICLR, and AISTATS

Dr. Ravid Shwartz-Ziv

Team Lead: Personalization & Adaptation

Dr. Ravid Shwartz-Ziv leads the Personalization & Adaptation Team at Wand AI, where he spearheads the development of cutting-edge large language models tailored to individual user needs. Concurrently, Dr. Shwartz-Ziv is an Assistant Professor and Faculty Fellow at NYU's Center for Data Science.

His work focuses on advancing the understanding and development of AI models, with particular emphasis on self-supervised learning, Bayesian deep networks, and information theory. Dr. Shwartz-Ziv's groundbreaking research explores the intricate connections between deep neural networks, information theory, and learning representations.

His contributions have garnered significant attention, with his work featured in prominent publications such as Quanta Magazine, Wired, ZDNET and many podcasts and talks. Dr. Shwartz-Ziv's expertise spans both academia and industry, with notable experience at Intel, Google AI, and various open-knowledge initiatives.