
Emanuel Hernández Castillo
Fractional CTO · Costa Rica
Tech leadership for companies building with AI.
Fractional CTO, technical interviews, architecture reviews and AI agents in production, for founders who need senior engineering judgment before they need a full-time hire.
- Intel
- Apple
- ex-CTO, Skillful AI
- Building potenci
Services
Four ways to work together.
Pick the one that matches the problem in front of you. If none does, the free call will tell us that in fifteen minutes.
Fractional CTO
Part-time technical leadership, for as long as you need a senior engineering voice at the table.
- For
- Startups and SMBs without a CTO, or with a team that has outgrown its first architecture.
- You get
- Architecture and roadmap, and build vs buy calls
- Hiring and running the engineering team
- Vendor and AI tooling decisions
- A technical voice with investors and clients
PricingCustom, scoped in discovery
Technical interviews & hiring
A hiring loop that finds engineers who can do the work, run by someone who has done it.
- For
- Teams hiring backend, AI or data engineers without a senior interviewer in house.
- You get
- Hiring loop and take-home design
- Technical interviews, run live
- A written evaluation per candidate
- Help closing the people you want
PricingCustom, scoped in discovery
Architecture review & due diligence
An outside read on a system or a startup before you commit money or a roadmap to it.
- For
- Founders checking their own stack, and investors looking at a technical company.
- You get
- Audit of the system, code and team
- Risks ranked by cost and likelihood
- A written report
- An action plan you can execute
PricingCustom, scoped in discovery
AI agents in production
Agents that answer, quote and book for real customers, designed, built and operated like the ones I run today.
- For
- Businesses that want an agent on WhatsApp or the web doing real work, not a demo.
- You get
- Agent design, tools and guardrails
- Build, regression bench and release
- Monitoring and a runbook
- Handover to your team, or ongoing operation
Pricing50-hour block, $10,000
Why me
Judgment from building, not from advising.
Trained where engineering is unforgiving.
GPU software at Intel, then data science for Apple TV+, the App Store and Apple Music. Large systems, where a decision gets measured before it ships.
I have sat in the CTO chair.
After leading ML teams, I co-founded Skillful AI, an enterprise agentic AI platform, and ran its engineering as CTO: architecture, hiring, clients and investors.
I still run AI in production.
Today I build potenci, a business OS with native AI agents, and operate agents for real businesses. One of them answers customers on WhatsApp in about 50 seconds; the human team took about 17 minutes.
Read the case study
How we work
Small first step. Then we decide.
Step 1 · Free, 15 min
Discovery
Fifteen minutes on the problem, the team and the deadline. If I’m not the right person, I’ll say so and point you to someone who is.
Step 2 · Paid, scoped
First engagement
A paid hour, an interview loop or a scoped review. You get something written at the end: notes, an evaluation or a report with the next actions.
Step 3 · If it works
Ongoing
We keep going: fractional CTO on a fixed weekly cadence, or a 50-hour block to ship a specific system.
Pricing
Clear prices where the scope is clear.
The discovery call is always free. Everything else is either a fixed price or a written proposal.
Consulting hour
$200per hour
An architecture question, a hiring decision, a vendor choice or a system that keeps breaking. Written notes and next actions after the call.
50-hour block
$10,00050 hours
One system shipped to production, usually an AI agent: built, tested, released and documented. Priority scheduling and weekly syncs.
Ongoing & scoped work
Customscoped in discovery
Fractional CTO, technical interviews and due diligence. Priced on scope and cadence once we know the team, the stage and the deadline, with a written proposal first.
NDA on request · blocks paid upfront
FAQ
Before you book a call.
The job a full-time CTO would do, on part of the week: own the architecture and the roadmap, make build vs buy and vendor calls, hire and lead the engineers, and be the technical voice with investors and clients. You get the judgment without the full-time salary or the equity of a co-founder.
A fixed cadence we agree on after discovery, sized to your stage: enough to make decisions and unblock the team, not so much that you pay for idle hours. We revisit it as the company grows or hires a full-time CTO.
Backend, AI and ML, and data engineers, from mid-level to lead. I design the loop with you, run the technical interviews, and send a written evaluation of each candidate with a clear recommendation.
A look at the system, the code, the infrastructure and the team, then a written report: what holds up, what is at risk, what it would cost to fix, and an action plan in priority order. Founders use it to plan; investors use it to decide.
Both. Leadership and reviews are mostly decisions and documents. AI agent work is hands-on: I design, build, ship and operate them, and the 50-hour block exists for that.
Costa Rica (UTC−6): working hours overlap with the US and Latin America, and with the afternoon in Europe. I work in Spanish and English. Everything is remote; NDA on request.
Start here
Fifteen minutes, free.
Tell me what you’re building and where it hurts. I read every form before the call, so the fifteen minutes go to your problem.
