Post-LLM research lab.
Dyssonance AI builds systems that form beliefs, revise them as evidence arrives, and continually reason under uncertainty.
Dr. Jennifer Prendki, PhD
Founder, CEO and Chief Scientist
Prendki trained as a particle physicist and earned her doctorate at the Sorbonne. She then spent nearly 20 years building and leading machine learning and data science teams, including as chief data scientist at Atlassian and in leadership roles at Walmart Labs (now Walmart Global Tech) and Figure Eight. She most recently led AI data at Google DeepMind.
At Dyssonance AI, she developed Continual Reasoning Under Uncertainty, a method that lets AI systems form and update beliefs as they deal with incomplete, unclear and changing information. Her time in AI at Google gave her direct insight into what works and what doesn't, so she built Dyssonance AI with a strong focus on ethics, privacy and security.
Where LLM-based systems break.
Enterprises are not seeing the returns they expected from LLM-based systems. The failure modes are consistent: token costs grow with every repeated context, coherence fades over long work, and answers degrade when information is incomplete or contradictory. That is how the world works.
Unlike LLMs, which predict the next word, Dyssonance AI uses a physics-based approach to intelligence that models thought and builds stateful, common-sense reasoning layers incentivized to find coherence and tensions (dissonances).
‘Agent harness’ is an oxymoron: we want AI to have agency, then constrain it because we don’t trust it. We’re changing that by building a different kind of belief-conditioned agency.
Continual reasoning under uncertainty
Conclusions stay provisional and get revised, decayed or strengthened as the evidence changes.
Dissonance detection
Contradictions are a key signal pushing work forward. The belief system is incentivized to find equilibrium between coherence and dissonance.
Belief-conditioned agency
Humans and agents interact with the same belief states. Agency conditioned by what the system believes, not constrained by a brittle harness.
