Techstrike.ai

APPLIED RESEARCH

Explore the questions
behind better AI.

How systems reason, retain information and make their behaviour easier to inspect. Published evidence and work in progress, clearly separated.

R / 01

AGICE

PREPRINT + PUBLIC EVIDENCE

Governed, evidence-native reasoning.

Verification, rollback and traceability, with public materials for inspecting the published experiments.

Publication
22 February 2026
Status
Preprint, not peer-reviewed
Open research dossier
R / 02

ZENGRAM

R&D INITIATIVE

Human-inspired AI memory.

An ongoing research direction, without unverified results or claims of released product integration.

Focus
AI memory
Status
Research and development
Explore the research direction

Evidence before generalisation.

Published experiments have defined methods and limits. A preprint is not peer review, and a research result is not a guarantee of production performance. Primary materials help readers inspect the work in context.

NEXT CONVERSATION

Have a question worth investigating?

Discuss a research collaboration or an engineering problem that needs careful investigation.

Discuss research