Designing AI experiences that amplify human capabilities.
Adagio for 69,996 starsRachmaninoff, Piano Concerto No. 2, II. Adagio sostenutoRecording: Musopen
I design the interface, build the system, and lead the team that ships it.
Making knowledge something you can work with.
Contextual tutoring, knowledge retrieval, and personalized study, in production. I lead a five-person team from product strategy and interface design through engineering and operations.
Debugging grounded in evidence.
Records what happened in the browser (actions, network, console, DOM changes) and turns the session into structured evidence that engineers and coding agents can inspect, reproduce, and debug.
Source on GitHubOne request, many coding agents.
Delegates engineering work across different coding agents through MCP, with controlled workflows and sessions that can be resumed.
Private. Request a demoCan a compact, traceable representation help AI navigate extensive sources?
Compact descriptors, relationships between topics, and traceable evidence selection, so a system can find the right passage without losing its connection to the original source.
The question came from building Haiku. The paper turns it into an implementation you can inspect.
Try the interactive demoSmall tools that run in your browser. Your files stay on your device.
A reusable skill for coordinating specialized coding agents with explicit ownership, bounded scopes, and acceptance criteria.
Source on GitHubMakes audio and video inspectable by AI agents through timestamped retrieval and focused media analysis.
Not public yet. Ask about itWhat I can help you build.
From the first product decision to the systems that keep it running, with the perspective of a founder who operates an AI product in production.
Tell me about your projectProduct discovery, roadmaps, UX/UI, interaction flows, onboarding, and technical specifications with clear acceptance criteria.
Contextual assistants, retrieval and RAG, semantic search, multi-agent systems, MCP integrations, and workflow automation. Model and provider choices informed by quality and cost.
Frontend and backend engineering, APIs, service boundaries, data modeling, and system design. Hands-on delivery with React, TypeScript, Python, Node.js, and SQL.
Serverless and cloud architecture, Cloudflare Workers, PostgreSQL, Linux, and Nginx. Production operations, performance, telemetry, infrastructure costs, and vendor evaluation.
Automated testing with Playwright and Vitest, evidence-based debugging, code review, release validation, authentication, and access control. Coding-agent environments with deterministic checks and human review.
Technical planning, work allocation, architecture decisions, stakeholder alignment, and engineering standards. Workshops and mentoring on AI-assisted development, harness engineering, and agent orchestration.
Rodrigo Garciaguirre
Founder, AI engineer, and researcher in San Salvador, El Salvador. At Haiku I connect product decisions with hands-on engineering and own the path from an idea to a running system.
45 minutes at the Codex Community Meetup, NubiWork, La Libertad. Coding agents, MCP, orchestration, deterministic validation, and human review.
Hackathon co-lead · November 14–15, 2026
Campus leader
Webmaster · University of El Salvador
University of El Salvador. Software engineering, databases, systems architecture, and infrastructure.
Bilingual Technical Assistant at Centro Cultural Salvadoreño Americano, 2021. TOEFL ITP, ETS, November 2021.
Let's build
something useful.
Thoughtful projects, technical challenges, and conversations about AI.
[email protected]