Partnerships and applied work
Start with a real question, not a logo wall.
Workshop practiceI build bounded research, architecture, workshop, and product systems around questions that matter to both sides. Specialist agents accelerate the work. I own the relationship, the judgment, and the result.
Workshop practiceOne consequential question
Scope the decision before the deliverable.
One bounded system
Make ownership, evidence, and limits explicit.
One inspectable result
Leave the team with something it can operate.
Relationship ledger
Every status is stated next to the claim.
Public proposals, independent platform briefs, and exploratory notes are different things. This page keeps them separate so the evidence can be judged without borrowed authority.
Independent platform alignment
Briefs about platforms visible in the work
These pages document public artifacts and, where a linked artifact supports it, Frank's own use or prior work. A proposed next step is not evidence of partnership, conversation, or endorsement.
Open notes
Opportunities with relationship status attached
Independent opportunity notes remain distinct from exploratory conversations and formal relationships. Each detail page states which one it is.
What can be inspected now
Current tools and public evidence.
Declared use, public artifacts, and relevant prior experience—not a claim that every company listed here has worked with FrankX.
Public build practice
Claude Code and Cursor for coding workflows; GitHub for public review; Vercel for the current frankx.ai deployment. Tool choice varies by task.
Model-learning surface
Claude, Gemini, Codex, ADK, A2A, and MCP appear in public workshops, guides, or research. Inclusion does not imply endorsement or production use.
Cloud context
Vercel is inspectable in the current deployment. Oracle Cloud reflects disclosed prior experience. No current OCI, GCP, or AWS workload count is claimed.
Accelerated-computing research
NVIDIA NIM and accelerator patterns appear as public research topics. No NVIDIA relationship, event work, or hands-on delivery is claimed.
A useful first conversation
Bring the decision your team cannot afford to get wrong.
In 30 minutes, we map the destination, the bottleneck, and whether a bounded partnership makes sense. No deck and no follow-up sequence unless we both want one.