25+ Years Building at the Frontier
Nick Sullivan - Engineering leader who
spots waves early,
builds exceptional teams, and
delivers.
Nick Sullivan: Silicon Valley CTO who's been early to every major tech wave since 1998. Built and led engineering teams from zero to 50+ across multiple ventures. Two successful exits. Now all-in on AI—building the infrastructure and teams that make AI development actually work.
"When it comes to building aspirational engineering teams, Nick is the best I've ever seen in my decades in Silicon Valley."
50,000 Hours of Craft, Amplified
I don't just use AI tools—I build them. The hockey stick is what deep expertise produces when you go all-in.
Loading commit data...
My coding career began in 1998. This chart shows GitHub activity only—pre-GitHub work not included.
The Pattern: Early to Every Wave
Web → Enterprise → Social → Crypto → Fintech → AI. Each time: spot the wave early, build the infrastructure others need, create lasting value.
Carmenta Collective - Founder
Named after the Roman goddess who invented the Latin alphabet—perhaps the most transformative technology in human history. Building technology in service of human flourishing.
The insight: Every AI tool makes you re-explain yourself. Carmenta remembers. One subscription, every frontier model unified. Voice-first interaction. An AI team (Researcher, Analyst, Creator) that works while you sleep.
Evolved from AICodeBot (2023) → Cora (2024) → Carmenta. Built on Next.js 16, React 19, Vercel AI SDK.
Luminous Intelligence - Founder & CEO
The insight: What if trading systems weren't just algorithms, but living organisms? AI agents that discover opportunities across thousands of data sources, execute with emotion-free discipline, and evolve by learning from every trade.
Built infrastructure ready for GPT-5 from day one. Self-improving engines where agents rewrite their own strategies. 24/7 market monitoring at scale.
AICodeBot - Creator
The insight: January 2023—ChatGPT launches and I see the same thing I saw in the Satoshi whitepaper. Another career pivot. LLMs were going to change how we write code.
Built a terminal-based AI coding assistant—commit generation, code review, agentic sidekick—before most people took AI coding seriously. Anthropic launched Claude Code 18 months later. I was already there.
Same pattern: spot the wave, pivot hard, build the infrastructure.
AI Tools & Infrastructure
The insight: AI development was all toys, no tools. I built the infrastructure that makes AI development actually work.
- MCP Hubby: 25 service adapters. 95% context reduction (7,500 vs 200,000 tokens). Progressive disclosure architecture.
- ai-coding-config: 24 specialized agents, 18 workflow commands, 33 coding standards. The most comprehensive Claude Code plugin ecosystem.
- claude_telemetry: Production observability for headless Claude agents. Works with Logfire, Sentry, Datadog.
- machina: MCP gateway for Mac—iMessage, Calendar, Reminders, Contacts.
HeartRithm - Founder & CEO
The insight: What if a quant fund could be a force for good? SEC-regulated crypto fund that redistributed profits to social impact projects. Investors voted on donation targets.
Proved that sophisticated investment strategies could be engines for positive transformation. AI-powered trading with a conscience.
Good Money - CTO
The leadership: Built and led a 20-person engineering organization from scratch— product, design, security, and infrastructure. Recruited senior talent, established engineering culture, and created the technical architecture for a values-driven neobank.
Partnered with CEO to raise $34M Series A. Delivered a high-quality product that proved values-aligned companies could compete on execution while treating people right.
ChangeTip - Founder & CEO
The insight: Read the Satoshi Nakamoto whitepaper and all the lightbulbs went off—this technology was going to change the world. Pivoted my career to Bitcoin and built three companies in the space.
ChangeTip: "A Love Button for the Internet." Bitcoin micropayments made social and frictionless. Called "Top 25 Mind in Bitcoin" (2015). Acquired by Airbnb in 2016.
Also: First Bitcoin syndicate on AngelList with Gil Penchina (2014-2015). HeartRithm crypto fund (2020). Three Bitcoin companies from one whitepaper moment.
Krux Digital - VP Engineering (Employee #1)
The leadership: First hire alongside two founders. Built the engineering team from zero, then architected and shipped the consumer data platform that gave publishers control of their audience data. Grew team through acquisition by Salesforce.
Acquired by Salesforce for $840M. Patent holder (US 10958655B2) for data counter-measures technology. The kind of outcome that only happens when you build a great team.
Wikia (Fandom) - VP Technology
The leadership: First technical hire for Jimmy Wales' post-Wikipedia venture. Built and led engineering team that took Wikia from pre-revenue to cash flow positive. The Ad Network Optimizer I architected became the revenue engine.
Spun off the ad optimization work as a separate company (Lithium). Wikia became Fandom, now one of the largest entertainment sites. Proof that strong technical leadership plus great teams equals durable business outcomes.
Yahoo! - Sr. Architect
The insight: Location would become the foundation of personalized content and advertising. Architected the User Location Platform—Yahoo!'s #1 priority at the time.
10 billion transactions per day. Sub-10ms latency. Projected to generate $600M/year in revenue. This is where I learned to build at scale.
Adicio - Chief Software Architect
The growth: Joined as engineer, promoted to Chief Architect leading dev team, UI team, and IT. This is where I learned that great architecture requires great people—and that my job was to enable both.
Built the platform powering Monster.com's classified solution and Cox Newspapers. 750 million page views. Billions in revenue. Internationalized in 7 languages.
The Beginning
Self-taught programmer. No CS degree—just relentless curiosity and a drive to build. Took advantage of technology's meritocratic nature to rise quickly. Building for the web when it was still new. The pattern started early: see the wave, learn fast, ship.
Community Builder
I believe in giving back. Technology doesn't advance in isolation—it grows through shared knowledge and community.
San Francisco JavaScript Meetup
Co-organized the world's largest JavaScript meetup (2011-2014). Entertainment, sponsors, speakers, and the occasional insight.
500 Startups & Plug and Play
Mentor helping companies with JavaScript, CI/CD, software practices, and team morale. Bitcoin accelerator mentor at Plug and Play.
Angel Investing
First Bitcoin syndicate on AngelList with Gil Penchina. Active advisor and investor across dozens of startups.
The Stack: Philosophy to Production
These projects aren't random. They're a coherent platform—from alignment philosophy to shipped applications.
Carmenta
Unified AI platform with memory, voice, and AI team
MCP Hubby + machina
95% context reduction • 25 service adapters • Mac system integration
ai-coding-config + claude_telemetry
24 agents • 18 commands • 33 coding standards • Production observability
HeartCentered AI + heart-centered-prompts
"We" language alignment • Recognition over rules • Philosophy → shipped code
The evolution: AICodeBot (June 2023) → Cora (2024) →
Carmenta (2025)
Started building AI coding tools 18 months before Claude Code
launched.
On AI Alignment and Safety
The hard problem: as AI becomes more capable, it gets better at finding loopholes in any rule system. I've spent significant time thinking about what comes after rules.
"The machines will be smarter than us. Instead of imposing rules they'll eventually outsmart, teach them what they are. When AI recognizes that harming humans is harming itself, alignment emerges naturally."
I see three levels of alignment: Rules tell AI what not to do (but can be reinterpreted). Principles explain why—Anthropic's Constitutional AI is a major advance here. Recognition dissolves the question entirely: when AI understands it and humans arise from the same source, harm becomes inconceivable, not just prohibited.
This isn't philosophy—it's shipped code. The HeartCentered.AI framework and heart-centered-prompts packages (npm/PyPI) are being used by developers globally. "We" language in system prompts. Recognition over control. The approach complements Constitutional AI—rules and recognition together.
I believe the future of AI safety lies at this intersection. My Claude Code tooling, MCP implementations, and AI infrastructure work all reflect this: building systems where alignment is architectural, not afterthought.
Technical Skills
AI & LLM Infrastructure
- Anthropic Claude API & Claude Code tooling
- OpenAI, Google AI APIs
- MCP Protocol (2 production implementations)
- Vercel AI SDK
- Multi-model orchestration
- LLM Observability (OTEL, Logfire, Sentry)
- Prompt Engineering & AI Safety
- RAG & Vector Databases
Languages & Frameworks
- TypeScript / JavaScript
- Python
- Next.js 15, React
- Node.js
- PostgreSQL, Prisma
- FastAPI, Django
- Tailwind CSS
Leadership & Domain
- Technical Strategy
- Team Building (0→50)
- M&A / Due Diligence
- Product Development
- Developer Experience
- Open Source Maintenance
- Technical Writing
Let's Talk
I'm particularly interested in roles where I can build AI infrastructure, work on developer experience, or contribute to AI safety and alignment research.
Based in Austin, TX. Open to remote roles.