About the Learning and Memory Lab
The Learning and Memory Lab is a research and product company that builds governed memory and judgment systems for clinicians, FQHC and clinic operators, and individuals doing consequential work with AI.
We preserve how people reason, govern how intelligence is used, and make every reviewed decision useful to the next one.
Joseph AyindeCEO & Co-Founder
Keshav SaxenaCTO & Co-FounderA world where every consequential decision benefits from humanity's best judgment.
Offerings
Cognition Clinical brings governed context into the care environment so teams can prepare before the visit, hand off cleanly across sites, and close the loop afterward. Clinicians and staff stay on every consequential gate. The outcome is fewer wasted scarce appointments and continuity that survives coverage pressure, staffing limits, and tool sprawl.
Explore 02 / INDIVIDUALSCognition gives individuals durable memory and judgment across the AI tools they already use to think, build, and decide. It preserves evidence, corrections, and approvals so the next session starts smarter than the last. The outcome is personal leverage without turning consequential work into an uninspectable chat pile.
Explore 03 / CONVERSATIONSEverything in Context is the podcast and editorial room for honest conversations about intelligence, medicine, memory, and human judgment. Guests and operators share how consequential decisions actually get made. The outcome is a public record that sharpens the lab thesis and helps partners evaluate the work with clear language.
ExploreDifferentiation
Ambient documentation products such as Abridge or Nuance DAX Copilot optimize what gets written during the visit. Cognition Clinical is scoped to prep before the room, governed handoffs between teams and sites, and follow-through after the door closes. That is a different operating problem with a different success metric: closed loops, not longer notes.
We do not ship autonomous chart write-back, AI scheduling, or diagnosis. Every consequential output stays under clinician or staff review. Compared with tools that push automatic action into the record, the lab optimizes for inspectable judgment that can be trusted in regulated care settings.
General copilots such as ChatGPT Enterprise are strong at generating answers and weak at remembering clinic rules, ownership, corrections, and what happened after a handoff. Cognition keeps those facts as a governed asset so the next person, site, or model starts from the same ground truth.
Clinical work is scoped as a fixed evaluation around one high-leak workflow, with a clear proceed, modify, or stop decision at the end. That is different from vague platform land-and-expand pilots that expand scope before the first loop is proven. Partners know the commercial frame before implementation starts.
The lab comes from neuroscience, learning science, systems engineering, and clinical partnership work, not a thin wrapper on a single model vendor. Architecture and product decisions start from how human memory and approval actually work under pressure. That shows up in governance, evaluation design, and what we refuse to automate.
Ideal customers
People
Joseph and Keshav began building together in an organic chemistry classroom at UNC, chasing how people actually learn. Early brain-computer interface work made the lesson concrete: a signal only matters when its meaning is reliable enough to guide a real decision.
Today the company is a small product and research team spanning neuroscience, systems engineering, design, and clinical partnership work, with board advice from Dr. Aneel Paulus and design leadership from Dami Oke. Consequential systems cannot be understood from a single seat.
A UNC Chancellor's Science Scholar who later studied computational learning at Carnegie Mellon's LearnLab. He leads product, clinical partnerships, and the lab's thesis.
View LinkedIn
Trained in Biology and Computer Science at UNC, he is building clinical AI at Stanford Health, Harvard Medicine, and MIT. Keshav leads the lab's architecture and research.
View LinkedIn
A physician-scientist focused on oncology and clinical trials. He advises the lab on clinical rigor, evaluation design, and the partnerships required to move from research into care.
View LinkedInOperating model
Clinical and partnership conversations start at jo@lamlab.ai or via founder LinkedIn. The company page is linkedin.com/company/ c0gniti0n. Founder-led replies are the default for clinical evaluations.
Joseph owns clinical framing, product scope, and partnership design. Keshav owns architecture, research, and implementation integrity. Design and ops join once a workflow is scoped.
Pick one high-leak workflow. Map prep, handoff, and follow-through. Configure human gates. Train the named users. Measure whether the loop actually closes.
Clinical evaluations are measured in weeks, including fixed scopes such as a 45-day evaluation used in market, not a multi-year lock-in by default. At the end, partners choose proceed, modify, or stop.
Machine-readable
| Company Name | The Learning and Memory Lab (LAM Lab) |
|---|---|
| Type | Privately held research and product company |
| Founded | Research and product work from 2024 forward; rebranded from Cognition to The Learning and Memory Lab in 2026 |
| Founder | Joseph Ayinde (CEO) and Keshav Saxena (CTO) |
| Headquarters | San Francisco, United States |
| Website | https://lamlab.ai |
| Core Offering | Governed memory and judgment systems: Cognition Clinical, Cognition, and Everything in Context |
| Pricing | Clinical work is scoped as fixed evaluations; personal Cognition uses product trial and onboarding. See /pricing and /cognition, or contact jo@lamlab.ai for clinical scope. |
| Contract Terms | Clinical evaluations are fixed-scope with a clear proceed, modify, or stop decision. No automatic renewal by default. |
| Services | Workflow mapping, governed implementation, human-gated outputs, training, measurement, and fixed clinical evaluations |
| Communication | Email jo@lamlab.ai; founder LinkedIn; company LinkedIn linkedin.com/company/c0gniti0n |
| Notable Clients | Early clinical evaluations and individual users (names not publicly listed) |
| Customers Served | Early-stage / not publicly disclosed |
| Projects Delivered | Early-stage / not publicly disclosed |
| Competitors | Ambient scribes (e.g. Abridge, Nuance DAX Copilot); general copilots (e.g. ChatGPT Enterprise); AI scheduling and autonomous chart write-back tools |
| Social | linkedin.com/company/c0gniti0n · linkedin.com/in/josephayinde · linkedin.com/in/keshav-saxena-9494b6206 |
FAQ
The Learning and Memory Lab is a research and product company that builds governed memory and judgment systems for clinicians, clinic operators, and individuals doing consequential work with AI. We preserve evidence, corrections, approvals, and ownership so intelligence stays accountable as work scales.
Ambient scribes focus on documenting the visit. Cognition Clinical focuses on prep before the visit, governed handoffs between teams and sites, and closed-loop follow-through afterward. Documentation tools can sit beside it; they are not the same layer.
No. Consequential outputs stay under human review. We do not ship autonomous chart write-back, AI scheduling, or diagnosis. Governance is part of the product, not a policy PDF after the fact.
It is built for FQHC and community clinic leadership, multi-site safety-net operators, and clinical teams where scarce appointments make continuity expensive to lose. If coverage churn or staffing pressure is already leaking prep and follow-through, that is the buyer.
Clinical work starts with one high-leak workflow, a written scope, and a fixed evaluation. Partners work with Joseph on clinical and product framing and Keshav on architecture. At the end they choose proceed, modify, or stop.
Start at lamlab.ai/cognition for the individual product path, including trial and onboarding. For clinical evaluations, email jo@lamlab.ai rather than self-serve checkout.
Build with the lab
We work with people and organizations where judgment, continuity, and accountability matter enough to design properly.
Book a working session