Chaos GoblinIndependent studio · Maryland · Est. 2026
⚠ THIS ORGANIZATION CONTAINS ENGINEERS
We build websites, software, experiments, and decision tools for people who need the thing to actually work—but would prefer not to die of boredom while getting there.
* Legal has asked us to stop making this claim.
What escapes the lab
CGDS sits somewhere between an engineering studio, a product lab, and the meeting where somebody says “that’s a terrible idea” right before we prototype it responsibly.
FIELD TECH // CLEARANCE: QUESTIONABLE
Custom design, rebuilds, responsive implementation, launch support, and ongoing care without template-factory energy.
Professional web development → 02 / ENGINEERINGAutomation, decision tools, internal utilities, and prototypes where spreadsheets have finally become an act of hostility.
Software & systems → 03 / LABHypothesis-driven research, failure testing, provenance, evidence, and explicit boundaries around what the goblin is allowed to touch.
Research & experimentation →Current specimens
Client work, internal products, and research systems. Status labels are honest. “In development” means exactly that.
Client services
Custom, maintainable websites built around real organizations instead of forcing organizations into templates. Released flagship build: Chesapeake Elite ↗.
Real client. Real launch. Extremely real QA.
Decision systems
A governed NFL daily-fantasy research platform for projections, identity resolution, lineup strategies, and prospective experiments.
Sports analytics with paperwork.
Music intelligence
Research into narrative playlist generation: intent, emotional arc, evidence, and why a song belongs where it belongs.
Taste is not a solved problem. Excellent.
Experimental resilience
A hypothesis-driven framework for safely testing how systems behave when dependencies, data, assumptions, and reality become uncooperative.
Failure, but make it governed.
Product development
A deliberately unnamed application currently moving through specification and prototype work. Details remain under wraps until there is something worth showing.
Yes, we know what it does. No, we are not telling you yet.
There are more ideas. This is not a threat. Probably.
The operating model
The brand is unhinged. The engineering is not. We use explicit assumptions, review gates, reproducible evidence, and human QA so the interesting ideas survive contact with reality.
CURRENT POLICY: “It compiled” is not user acceptance testing.
What are we actually trying to fix? What counts as success? What would prove the idea wrong?
No fake certainty. No mystery architecture. No unnecessary platform empire before the first useful result exists.
If a normal person cannot understand or use it, technical success has not yet occurred.