Engram

Company

Architecting cognition

We build full-stack intelligence systems -- integrated architectures that combine models, memory, reasoning, and execution into coherent systems that operate over time.

Founded 2024 · San Francisco · Melbourne · London

The Engelbart Thesis

Augmenting human intelligence

In 1962, Douglas Engelbart proposed that computing's purpose was not automation but augmentation -- increasing humanity's capability to approach complex problems.

We continue that work, building cognitive infrastructure for researchers and engineers at the frontier of their fields.

Full-stack intelligence

Five layers, designed together

Each layer amplifies the one below it. Because we build them together, we can optimise across the boundaries that other approaches treat as API walls.

05

Decision Engine

ProvenanceDeterministic replayLineage tracing
04

Agents

MonadTask orchestrationTool runtime
03

Memory

Engram-LocusPersistent state across sessions
02

Models

Engram-VQMulti-Doc ReasonerSpecialised APIs
01

Hardware

Custom compute for inference and training

Applied by design

Research shaped by deployment

Our systems are developed and tested in real environments where correctness, auditability, and reliability matter.

Deployment constraints that don't appear in benchmarks inform our research. Every real-world engagement surfaces problems that advance the systems themselves.

Speed

R&D throughput -- question to validated answer

Depth

Reasoning depth across long horizons

Reliability

Consistency under complexity

Auditability

Traceability without friction

Locations

Founded 2024. Four offices across three continents.

San Francisco skyline

San Francisco

United States

Headquarters
Melbourne skyline

Melbourne

Australia

R&D
Los Angeles skyline

Los Angeles

United States

Office
London skyline

London

United Kingdom

Opening 2026

Work with us

Deploy our systems, collaborate on research, or join the team.