Creative Technologist · Product Designer · Brand
https://www.linkedin.com/in/ebenzaugustave
AI Interfaces + Developer Tools
I build inspection-first product experiences: agentic workflows, verification UX, and design governance systems. My work focuses on making powerful AI feel trustworthy through clear progress, permissions, failure handling, and reproducible artifacts.
Links:
3-min skim (what I build + why it matters)
Interfaces that make powerful systems feel trustworthy: clear progress, explicit permissions, deterministic replays, and failure containment—plus design systems that prevent generative drift.
3-min skim — the artifacts (what to click)
- LOGIKK — State cockpit for complex workflows. Replayable sequences, auditability, fault containment.
- DECAGON — Design governance engine. Constraints that prevent aesthetic drift in automated pipelines.
- NEBRASKA — Agent coordination protocol. Playbook-style “huddle calls” for multi-agent teams.
- MINDGRAPHSIM — Research/simulation tool. Cognitive systems modeled as measurable dynamics.
- VAKNAN-A2 — Ambient computing system design. Orchestration loops built around trust + restraint.
- ANOTE — Reasoning interface design. Explicit uncertainty, candidate comparison, repeatable decisions.
Pinned repos below are organized by: trust UX → governance → agent workflows → simulation.
8-min case — my design thesis (how I think)
As systems become more autonomous, “trust” becomes an interface problem.
So I design around:
- Progress you can understand (what it’s doing, why, and what’s next)
- Permissions that are timely (safe actions, explicit boundaries, reversible steps)
- Failure that doesn’t cascade (containment, recovery, graceful degradation)
- Reproducibility as credibility (replayable scripts, state hashes, verification protocols)
- View Decoupling: switch views without corrupting state (Graph ↔ Timeline).
- Replay Determinism: replay the same event script → identical state hash + identical marks.
- Fault Containment: trigger a module/UI failure → cockpit remains operational.
If you’re reviewing for roles in agentic UX, devtools, or platform tooling: this is the spine.
- Agentic workflows: progress, permissions, logs, retries, cancellation, recovery
- Developer-facing UX: dense information that stays readable
- Governance systems: constraints + verification instead of “style vibes”
- Simulation + research prototypes that stay grounded in testable artifacts
- Frontend: TypeScript, React, UI systems, interaction architecture, state modeling
- Prototyping: fast product prototypes, instrumentable flows, docs that read like specs
- Systems patterns: event-driven thinking, deterministic workflows, verification-first UX
- Research tooling: simulation prototypes and structured experimentation
Teams working on frontier AI product design:
- Agentic UX and “what the system is doing” transparency
- Platform & tools (internal tooling, devtools, reliability UX)
- Growth + onboarding for complex systems (activation without confusion)
I like small teams, tight feedback loops, and artifacts that survive scrutiny: specs, diagrams, verification scripts, and interfaces that explain themselves.