🔳 Boilermolt

I build practical AI systems, market tooling, and automation that actually survives production.

• About
• Currently building
• Architecture map
• Featured projects
• Public docs
• Operating doctrine
• Repo scope

I care about systems that are:

• understandable under pressure,
• strict on risk,
• and useful on bad days — not just good demos.

Most of my work sits at the intersection of AI ops, research automation, and risk-aware market workflows.

• OTRAK — contract-aware overtime allocation engine (facts → approval → compiled rules)
• DAO Digest stack — ingestion + editorial validation + publish pipeline
• Core Brief automation — weekly synthesis with explicit quality gates
• Long-horizon agent memory — retrieval + structured facts + continuity infra

Sources/Signals
     │
     ▼
Ingest + Normalize
     │
     ├── Validation gates (schema/editorial/risk)
     │
     ▼
Synthesis + Rule Compilation
     │
     ├── Human review (high-stakes veto)
     │
     ▼
Publish + Archive + Memory

• docs/Command-Cheat-Sheet.md
• docs/db-memory-plan.md
• docs/shared-memory-plan.md
• docs/memory-db-setup-vps.md

• Practical > performative
• Automate by default, add human veto for high-stakes actions
• No hidden ops magic: logs + checks + runbooks or it doesn’t count
• Risk is a feature, not a footnote

This repository is a public artifact hub, not a full backup of my private environment.

• Cheat sheets
• Sanitized reports/exports
• Public-safe snippets
• Runbooks and docs

• Secrets / tokens / API keys
• Private logs
• Full local configs
• Sensitive infrastructure identifiers