Lab / experiments and ideas

Lab

Experiments. Prototypes. Questions still open.

The Lab is where I work through questions around agents, retrieval, tool protocols, agentic commerce, and ML/data work. Some connect directly to my day-to-day engineering workflow; others are earlier technical investigations.

A dashed exploration boundary containing a central shared-context core, wired to three labelled clusters: agents with context and review, retrieval with index and context, and protocol with boundary and tools. Every connector is drawn as in flight rather than settled.

How I use the Lab

  • Technical notes and working models
  • Questions tested through small experiments
  • A record of what I am learning and what remains uncertain

Lab notes

Current explorations

Each one has a working system shape and a set of questions still unresolved.

Compare with shipped systems
EXP / 01
AGENTCONTEXTTASKEXECUTEVERIFYRETRY
Exploring

Agentic Development

Practical ways to make coding agents useful inside real engineering constraints.

  • Agents
  • Context
  • Verification
Open notes
EXP / 02
SOURCESINDEXRETRIEVECONTEXTEVALUATE
Exploring

Local RAG

Exploring private, inspectable retrieval workflows for local technical knowledge.

  • Retrieval
  • Local data
  • Evaluation
Open notes
EXP / 03
PROTOCOL BOUNDARYCLIENTMODELTOOLSRESOURCESSERVERSREQUEST
Exploring

MCP & Tooling

Exploring small, explicit tool interfaces that connect agents to useful systems.

  • MCP
  • Tools
  • Boundaries
Open notes
EXP / 04
AGENTCATALOGIDENTITYCHECKOUTCOMMERCE PLATFORMMAPPING
Exploring

Universal Commerce Protocol

Exploring how open commerce protocols could let agents act through standardized capabilities instead of platform-specific integrations.

  • Agentic commerce
  • Protocols
  • Interoperability
Open notes
EXP / 05
TASKATTEMPTVERIFYDONELESSONSPASSIMPROVE
Exploring

Self-Improving AI Agents

Exploring how agents improve across attempts through execution feedback, verification, memory, and evaluation.

  • Verification
  • Feedback
  • Evaluation
Open notes
EXP / 06
RAWCLEANLABELTRAINVALTEST
Exploring

ML & Datasets

Exploring how raw data becomes a reliable machine-learning dataset, and what has to be true before training starts.

  • Datasets
  • Quality control
  • Reproducibility
Open notes