About

Tunjay Seyidali is the founder of Mnemoria.

He built it after noticing something most people never stop to examine: how often we act on beliefs we’ve quietly stopped holding. We tell people we’re fine with something we’re not. We describe ourselves one way while our actual behavior, tracked over months, tells a different story. Nothing catches this drift, because nothing remembers long enough to see it.

Seyidali is a final-year Computer Science student at King’s College London. His background sits at the intersection of engineering and analysis: he has built production backend systems within the banking sector, worked in quantitative strategy, developing optimization models and stress-testing pricing systems for industrial clients. He acted as a Youth Delegate to the United Nations Economic and Social Council and at the World Bank Youth Summit, authoring national policy positions on digital inclusion and economic transformation.

Mnemoria is his first product conceived, architected, and shipped entirely independently, from the underlying AI systems to the infrastructure to the design.

The technical premise is straightforward to state and difficult to build well. Most AI assistants can now recall a fact you’ve told them. Mnemoria goes further: it extracts what you actually believe from how you talk over time, tracks how those beliefs shift, and forms a real, synthesized perspective on you, the way a close friend would, rather than simply retrieving stored facts on request. In testing, the system has independently connected a user’s stated restlessness at a new job to an offhand comment about wanting to day-trade crypto, unprompted, calling it “the exit ramp you’re already building.” It has correctly recalled and applied a struggle a user described weeks earlier, after the entire visible conversation history was deliberately erased, proving the understanding lives in structured memory, not just chat transcripts.

The project sits at a genuine intersection of consumer product and open research question. If a system adapts continuously to one person over time, does it risk reinforcing that person’s worst patterns rather than helping them see clearly, the way a sycophantic friend might? Mnemoria’s answer is a specific design principle, not a slogan: adapt in tone, never in values. The system tracks how someone communicates and adjusts its own register accordingly, warmer or more direct, longer or shorter, matching the person. But its honesty, its willingness to name a contradiction, and its core character never bend to what might be easiest to hear.

Mnemoria is named for the Greek mneme, active recollection rather than passive storage, and draws its visual and philosophical language from an older idea: that the pineal gland was once believed to be the seat of inner vision, the point where mind meets matter. Mnemoria is built on that same premise, that real understanding of a person requires a place where memory and honesty meet, and stay.