Gloqo AI
Gloqo AI

Gloqo AI

Field notes on agentic AI systems.

We work quietly on the layer around frontier models: context engineering, harness engineering, retrieval, tools, memory, evaluation, and the deployment discipline that makes agents useful in production.

Operating notes
Private by default. Public when useful.
FocusAgentic AI systems, context engineering, harness design
ModeForward deployed, eval-led, instrumented in production
Public surfaceField notes, visual labs, operating patterns
$ cat publication_policy.md
Client work stays confidential.
Public work is limited to field notes,
visual labs, and operating patterns.

$ cat where_we_operate.md
Singapore, Warsaw, Melbourne, Sydney.

What We Publish

The site should feel like a trusted engineering notebook: useful, selective, and calm. The work tracks how frontier labs and product teams turn capable models into agents that can inspect context, operate tools, verify their own work, and stay inside clear boundaries.

Systems
Context engineering

The right instructions, retrieval, memory, tools, and schemas at the right moment.

Runtime
Harness engineering

Model-native execution loops, tool choice, sandboxing, self-checks, and traces.

Evaluation
Agent operations

Task evals, guardrails, permissions, handoffs, latency, and cost discipline.

Latest Field Notes

The newest research notes from the library.