{{ p.title }}
{{ p.body }}
Relational intelligence for real systems
AI that understands what connects, what changes, and what matters next.
We build learning systems that model relationships between signals, hold context over time, and support decisions in clinical, wearable, and population-scale settings.
01 / Platform
world models for work that really matters
We connect fragmented signals across people, devices, records, and conversations into representations that support grounded reasoning and calibrated action.
Illustrative system view
From signals to situated action.
{{ d.step }}
{{ d.title }}
{{ d.body }}
observation → representation → decision → learning
02 / Applications
where context becomes operational advantage.
We work where the quality of an action depends on understanding the surrounding system, not merely answering a single question.
{{ a.n }}
{{ a.title }}
{{ a.body }}
03 / Research
selected work, newest first.
Papers, talks, and challenge reports across affective computing, wearable AI, computational biology, clinical forecasting, epidemiology, and audio.
{{ r.domain }}
{{ r.title }}
{{ r.venue }}
{{ r.status }}
04 / Programs
programs open to partners.
Work in progress that organizations can join directly, rather than wait to read about. Each one takes a small number of partners at a time.
{{ p.type }}
{{ p.title }}
{{ p.meta }}
{{ p.cta }}05 / Achievements
competitions and exhibitions.
Placements and shows, filed separately from papers because they are judged on different terms.
Meta Wearable AI Grand Challenge @ ECCV 2026
Runner-up and third place in EgoConv.
Multi-turn egocentric conversation, judged in two model-size divisions. Our 1.3B model placed in the Small track; a larger variant placed in the Large track.
Public leaderboard →
{{ p.place }}
{{ p.track }}
MadArts · 2025
Sensation of Agency: “Feeling Control”
Audiovisual interactive exhibition.
See the work at brwn.art →06 / Team
a small team, led from the research side.
{{ t.name }}
{{ t.focus }}
{{ t.bio }}
07 / Company
a research company built for useful intelligence.
We believe capable systems should understand more of the situation around a decision, make their uncertainty legible, and expand rather than obscure human agency.
“The future of AI is not a better autocomplete. It is better orientation inside a living, changing world.”
Everyday Intelligence is an independent research and product studio. The work spans clinical prediction, wearable and ambient sensing, population health, affective computing, and enterprise operations, and it is expanding into new domains with the partners who bring them.
01Build from the structure of the problem, not from a fashionable interface.
02Report the metric that can be gamed alongside the one that can't.
03Design systems people can inspect, guide, and meaningfully overrule.
Notes
Working notes on methods, failures, and what the metric hides.
Short technical writing between papers. Roughly monthly. Where a note has a paper behind it, both are linked.
Contact
Bring a difficult, real-world problem.
We take on client engagements, pilots, and build work, and we are open to research collaborations alongside them. Tell us what you are trying to predict, what you already tried, and what a good answer would change.
hello@everydayintelligence.org →