Be part of Fintella Labs.

What we're working on now.

• Knowing when a read on someone is actually right

  When a person recognizes themselves in a description, it means the description sounds right, which is not the same as being right. We work on how to really check a behavioral read when there is no clean outside answer to compare it to.

• Catching when someone's life is shifting, as it happens

  Looking back, a change is obvious. Spotting it live, and knowing how sure to be, with no labels to learn from and a person who keeps moving, is the hard part.

• Separating what the data really shows from what we are reading into it

  We only ever see the choices people actually made, so pulling cause and effect out of that is genuinely hard, and having richer data does not make it any easier.

• Reading emotion from data that was never meant to measure it

  Whether you can reliably tell how someone feels from behavioral traces is still an open question, and we are testing how far it actually goes.

• What consent means when a system tells you something new about yourself

  Behavioral inference creates information that did not exist before, and the legal and ethical rules around it are still unsettled. We work right in the middle of that.

We would love to hear from people who work in these areas: computational behavioral scientists, psychometricians who care about validity, behavioral economists working with revealed preference, privacy researchers who know reconstruction attacks, ML people who have worked with longitudinal data, causal inference researchers, and engineers who like building while the ideas are still moving.