1 Senior AI Application Engineer - SD, Remote: Colombia - Costa Rica, Fulltime

- This position is open to candidates located in Colombia or Costa Rica only -
You'll build new product features on top of LLMs, including agents, retrieval systems, and the workflows around them, and you'll use AI tooling heavily to do it. This is a greenfield role on a small, fast-moving team. Requirements arrive loosely defined, you help define them, and you own what you ship from prototype to production. Our first major milestone lands 60 days out, so you should expect to be contributing code in week one and shipping something real well before your first quarter is up.
What you'll do
Design and ship user-facing features backed by LLMs, including evaluation, failure handling, and cost management
Make architectural decisions and defend them, weighing tradeoffs across speed, cost, complexity, and risk
Build systems that stay fast and available as usage grows, with sensible latency budgets, graceful degradation, and no single points of failure
Handle PII and financial data responsibly, thinking through what data flows where, what reaches a model provider, and what the failure modes look like
Build the automation and data pipelines those features depend on
Own deployment and operation of your services
Help set the team's direction on AI tooling and approach
Required Qualifications
Senior-level software engineering experience. You can design a system, review someone else's code, and debug something you didn't write
Architectural judgment. You think in systems rather than tickets, and you can explain why you chose one approach over another to both engineers and non-engineers
Experience assessing risk in environments handling PII or financial data, and a working sense of where the real exposure sits versus where it only looks scary
Track record building highly available, scalable, performant systems, and the instincts to know when that matters and when it's premature
Ships production code with AI coding tools such as Cursor or Claude Code as a normal part of the workflow, not just experimentation
Has built and shipped LLM-backed features to real users, such as agents or RAG pipelines, including the unglamorous parts
Practical judgment about where AI helps and where it doesn't, grounded in things you've actually built
Comfortable owning deploys end-to-end on a PaaS like Fly.io or Render, and able to find your way around AWS or GCP when a project needs managed services
Builds automation and data workflows to support the above
Moves fast under real deadlines and works well with ambiguous requirements
Motivated by building new features rather than supporting existing systems
Preferred Qualifications
AI-assisted code review workflows
Evals, fine-tuning, or model training experience

Work arrangement
Yes