A \ V · Radical Transparency

How to Work
With Me

What makes collaboration work. What drains it. How I think, communicate, and where I add the most value. Honest, not polished.

01 · Working Style
🔭Systems first, solutions second

I spend more time understanding the system than jumping to solutions. A well-framed problem is 70% of the solution. I'll ask "why are we solving this" before "how."

📐Precision in language

Words like "agent," "intelligent," and "scalable" mean specific things to me. I'll push back on imprecise framing — not to be difficult, but because vague language produces vague systems.

🔄Iterative over speculative

I'd rather build something real and learn from it than design the perfect system on paper. Architecture documents that never ship are fiction. Get something in front of users.

🎯Outcome-oriented

I care about what the system does for real people, not the elegance of its internals. Beautiful architecture that doesn't deliver value is an expensive hobby.

📣Direct communicator

I say what I think, diplomatically but clearly. If I think a technical direction is wrong, I'll say so with reasons. I expect the same — challenge me, cite your reasoning.

🌐Platform-agnostic conviction

No vendor loyalty. AWS, GCP, Azure, OpenAI, Anthropic — the right tool for the problem. I distrust architects who recommend the same stack regardless of context.

02 · Communication
✓ Works well
Async written communication — time to think before responding
Concrete problems with context — what you tried, what failed
Disagreement with evidence — pushback that makes me think harder
Short meetings with clear objectives and a decision to make
Diagrams, ADRs, written specs over verbal architecture sessions
Honest status including blockers — early signal is always better
✕ Friction points
Consensus-seeking without a clear decision-maker
Meetings to share information that could be an email
Vague requirements without acceptance criteria
Technology choices driven by trend rather than problem fit
Post-hoc documentation of systems never designed
Optimism theater — saying things are fine when they're not
03 · Energy

Energizes me

Hard technical problemsNovel agentic architecturesCross-domain thinkingTeaching & mentorshipHealthcare AI with real patient impactBuilding things that didn't exist beforeInterpretability & AI safety research

Drains me

Performative meetingsBikesheddingUndocumented tribal knowledgeVendor-driven architectureAnalysis paralysisRewrites without diagnosis
04 · Working Principles
1
Context before opinion

I try to fully understand a situation before forming a view. Asking basic questions is not lack of experience — it's refusing to optimize for the wrong thing.

2
Strong opinions, weakly held

I'll defend a position clearly but update it when presented with better evidence. The goal is the right answer, not being right.

3
Write things down

If it wasn't written, it didn't happen. Architecture decisions, trade-offs, post-mortems — these create the institutional knowledge that lets teams move fast without breaking things.

4
Measure what matters

Not everything measurable matters, and not everything that matters is measurable. I'm suspicious of metrics that optimize for proxy variables and miss the actual outcome.

5
Protect the team's ability to learn

Blame cultures destroy the information flows that prevent the next incident. Psychological safety is how you build systems that surface failures before they become catastrophes.

05 · Best Fit
🏗️
Architecture Advisory

Agentic AI system design, LLM orchestration patterns, multi-cloud AI platform strategy.

🏥
Healthcare AI

Revenue cycle automation, clinical decision support, prior authorization, compliance-aware AI.

🎓
Team Upskilling

Workshops on agentic AI, MCP integration, prompt engineering, MLOps maturity.

If this resonates

Let's build something together

The best collaborations start with a clear problem and honest expectations.

Get in touch →