AI Developer Relations Engineer

Felix Technologies, Inc.•Miami, FL

12h•$164,000 - $182,000•Hybrid

About The Position

Félix is building an AI-powered financial companion for Latinos in the US, beginning with cross-border money movement and expanding into credit, savings, and wallet services. The company describes itself as a hyper-growth Series C company backed by more than $300 million in funding. The Applied AI Engineer will help the AI team and colleagues across Félix use AI effectively in their daily work. The role combines hands-on engineering with technical office hours, mentoring, and role-specific learning programs. The engineer will build and debug agents and integrations, support colleagues’ builds, and turn successful approaches into reusable frameworks and learning resources. The goal is to help teams use AI independently, resolve technical blockers, and maintain their AI workflows. The role works with team leads and Internal Talent to identify skill gaps, track applied learning, and follow up after training. It develops expertise in Maestro and other approved platforms, partners with the AI Solutions Manager on shared practices, and engages FDE, AI Enterprise, Data, Security, and platform specialists when deeper support is needed. Production systems remain the responsibility of FDEs and specialist engineers.

Requirements

Hands-on experience building LLM applications, agents, or automations for real users, including experience with failures and improvements.
Strong Python, APIs, authentication, Git, testing, and debugging skills.
Practical use of at least one LLM SDK, such as OpenAI or Google.
Experience with tool calling, multi-step workflows, state, structured outputs, human review, and failure handling.
Ability to build useful evaluations and explain quality, latency, cost, and data safeguards.
Experience mentoring builders and mixed technical and business audiences through live builds and debugging, with follow-through after training.
Ability to read logs and traces, isolate failures, and create reproducible examples.
Ability to turn ambiguous requests into testable hypotheses, recognize when to escalate, and close the loop.
Must complete a live pairing or debugging exercise during the interview.

Nice To Haves

Practical SQL and CI/CD experience.
Experience with RAG and vector databases.
Experience building maintainable starter kits or reusable components.
Experience supporting operations teams.

Responsibilities

Build and improve reference agents, tool integrations, and workflows using Python, LLM SDKs, and APIs.
Pair with colleagues to scope practical builds, review code, and debug prompts, retrieval, tool calls, state, permissions, and failures.
Run regular technical office hours, triage requests, and track issues through resolution.
Escalate complex integration, data, security, or production problems with reproducible examples, logs, and attempted fixes.
Turn recurring technical issues into resources.
Create starter kits, tested agents, integration patterns, and troubleshooting guides based on shared standards.
Provide examples and guidance for tool calling, data access, evaluations, human review, and failure handling.
Maintain clear documentation, versioning, permissions, and ownership on approved platforms, including Maestro.
Own technical content and exercises for hands-on bootcamps and learning tracks, and teach through live builds and debugging.
Mentor colleagues through guided builds and code reviews until they can maintain their own work.
Maintain working expertise in Maestro, other approved tools, and data safeguards.
Test new platform capabilities on real tasks, share practical guidance, and bring recurring gaps to platform owners.
Track independent task completion, blocker resolution, resource reuse, and sustained adoption.
Work with team leads and Internal Talent to identify skill gaps, track applied learning, and follow up after training.
Partner with the AI Solutions Manager on shared practices and engage FDE, AI Enterprise, Data, Security, and platform specialists when deeper support is needed.