Loom AI Labs GitHub

Loom AI Labs

Independent applied engineering

Applied Research

Engineering investigations into reliable AI enablement for Java applications—covering live data, retrieval evidence, governed actions, privacy and tenant boundaries.

Browse investigations Explore live experiments

Implementation-linked · Limitations stated · Independently developed

Research lifecycle from question through limitations
Featured investigation Reproducible demo

Keeping Application Data and AI Evidence Aligned

Research question:
How can an application make entity-to-index consistency visible and recoverable as source records change?
Method summary:
Model create, update and delete as explicit entity lifecycle operations.
Limitation:
The live proof uses a single application instance and a compact synthetic dataset.
Read investigation Reproduce experiment
Application data and AI evidence revisions aligned through a lifecycle

Architecture study

Explicit Application Context Instead of Model Guessing

Research question: Which request context must be resolved by application code before AI orchestration begins?

Context and grounding
Read note Related experiment

Reference implementation

A Governed Lifecycle for AI-Proposed Actions

Research question: How can an application accept useful model-proposed parameters without weakening trusted resource and execution boundaries?

Actions and governance
Read note Related experiment

Security investigation

Tenant Identity as Orchestration Context

Research question: How should tenant identity cross retrieval and action boundaries without becoming prompt text?

Privacy and identity
Read note Related experiment

Security investigation

Privacy-Aware Context Preparation for RAG

Research question: Where should detection, redaction and encrypted-original handling occur before retrieval or generation?

Privacy and identity
Read note Related experiment
Design rationaleExploratory ideas and reasoning
Implemented prototypeCode and working implementation
Reproducible demoExperiments others can run
Measured evaluationEmpirical results and analysis

Only published evidence levels are claimed.