AI Integration
Claude and LLM features that actually belong in your product. RAG, agentic workflows, internal tools, content pipelines. Built with the APIs, not against them.
CONSULTING · SOFTWARE · AI
Technical consulting for international teams hiring senior LatAm engineering, and for Costa Rican companies that need to modernize.
Flexible scope. Modern stack. Senior judgment. No layers between you and the work.
Claude and LLM features that actually belong in your product. RAG, agentic workflows, internal tools, content pipelines. Built with the APIs, not against them.
CI/CD, test infrastructure, QA tooling, data pipelines. Fifteen years shipping internal platforms at scale. Your team stops firefighting and gets back to building.
Full-stack applications, APIs, backends. Python, TypeScript, Java, C/C++. Built to run in production for years, not to look good in a demo.
Architecture review, stack decisions, technical hiring. A senior opinion that isn't trying to sell you a specific stack.
Testing infrastructure where LLMs do the grunt work and engineers validate the verdict. Coverage analysis, flaky-test triage, log investigation. Faster than manual QA, safer than fully autonomous AI. The testing discipline Costa Rica is known for, applied with modern tools.
Short essays, competitive teardowns of the Costa Rican tech landscape, and honest writeups of what worked and what didn't. If you want new posts in your inbox, email me.
READ NOTES→I'm Sergio Barahona. Senior software engineer with fifteen years building production systems, including deep experience in QA tooling and test automation at enterprise scale. Python, TypeScript, Java, C/C++, Docker, Kubernetes, CI/CD, and the judgment to know which of those matters for your problem. Costa Rican, bilingual, pragmatic. I started Baher to work on my own terms: selective projects, direct relationships, and the honesty to tell a client when an idea won't work.
What Sergio delivers isn't code, it's clarity. He walked into a vendor decision we'd been debating for months and reframed it in thirty minutes. We realized we needed a QA strategy, not a QA platform. That conversation alone saved us well into six figures.
A short email works best. Describe the shape of the problem and I'll reply with whether I'm the right person for it.