Senior Full Stack AI Engineer

Role Overview

We're hiring two senior full-stack engineers to build the next generation of Ombud's agentic AI platform. The work splits across two domains: building the agentic engine itself and the ML data engineering that supports it.

What You Will Do

You will ship production code, own systems end-to-end, and operate without a layer of engineering management between you and the product direction. You will work directly with the CEO on architectural decisions and directly with the platform engineer on production deployment.

Why It Might Be a Fit

This is the engineering team that builds the actual product behind the agentic enterprise era. We use Claude as a teammate, not a feature checkbox. We deploy frequently, ship real customer value, and trust our engineers to operate as senior partners — not as cogs in a sprint.

Requirements

• 6+ years of professional full-stack software engineering experience, with demonstrated production system ownership
• Deep Python fluency
• JavaScript / TypeScript / React competence
• Hands-on production experience integrating LLMs into product (Anthropic, OpenAI, Google)
• Working knowledge of RAG architectures, embeddings, vector databases, and the trade-offs between retrieval and context-caching approaches
• Fluency with Claude Code or similar AI-augmented development workflows
• Strong intuition for system design
• Comfort operating in a small team without a layer of engineering management
• Willingness to be in-office Tuesday through Thursday in Denver

Benefits

• Production operations
• Code review and technical mentorship within a small, senior engineering team
• Self-service infrastructure
• Full-stack feature delivery across Python, Node.js, React, PostgreSQL, and Elasticsearch
• Production agent orchestration, tool design, multi-step reasoning workflows
• RAG and CAG architecture: embeddings, retrieval, re-ranking, caching strategies
• LLM integration with Anthropic Claude (primary), with multi-model routing where it makes sense
• Evaluation pipelines: how we measure response quality, regression-test prompts, and ship LLM-dependent features with confidence