Macomber Tech

Independent Research & Development

Independent Research & Development

Practical technology investigation.

Research into software systems, automation, platform engineering, artificial intelligence, and emerging technologies.

Macomber Tech explores new technologies by building with them. Projects begin as technical questions, then evolve into notes, tools, prototypes, reference architectures, or lessons learned.

Research Statement

Projects begin as technical questions. Some become notes. Some become tools. Some become prototypes. Some become dead ends. All produce knowledge.

Featured Investigations

Current investigations and durable field reports.

active AI systems

Local LLM Infrastructure for Practical Software Development

Evaluating whether local models can participate usefully in real software development workflows without giving up privacy, speed, or operational sanity.

Can local models contribute meaningfully to planning, implementation, and review inside practical engineering workflows?

active Observability

Observability as a Queryable System

Treating operational telemetry as a system that should answer questions directly, not only populate dashboards after the fact.

What changes when observability data is designed for direct technical investigation instead of passive dashboard consumption?

complete AI systems

GLM-4.7-Flash on DGX Spark with native upstream vLLM

Field report on getting GLM-4.7-Flash running natively on DGX Spark with upstream vLLM, including dead ends, compatibility traps, and the final known-good stack.

What stack reliably serves GLM-4.7-Flash on DGX Spark using upstream vLLM instead of helper repositories or vendor-specific wrappers?

Current Research Areas

  • AI agents and agentic software development workflows
  • Local LLM infrastructure and private AI systems
  • Platform engineering and developer enablement
  • Reliability, observability, and operational excellence
  • Systems automation and workflow optimization
  • Knowledge capture, retrieval, and organizational learning
  • Software architecture and distributed systems

Recent Notes

Thinking Before Delegating to AI

AI output usually reflects missing structure in the request. Clearer thinking before delegation improves results more than repeated prompting.

Selected Artifacts

Tools, prototypes, and public outputs from research work.

tool published

Mimir

Interactive public artifact available as a standalone tool surface.

Approach

The objective is not to chase technology trends. The objective is to determine what is useful, what is practical, what can be maintained, and how emerging technologies can be applied effectively to real problems.