AI Systems Lead

Waterstone Human Capital; Salt Lake City (UT); Presencial; USD 140.000 - 190.000; Jornada completa

Ventajas ofrecidas por este puesto de trabajo

Equity available

Descripción de la vacante

Waterstone Human Capital is seeking a senior AI systems engineer to own and evolve an AI operating system that tracks commitments across the company. You’ll maintain health checks, run logs and alerts, and regularly review whether each agent remains useful.

You’ll build new agents from team specifications and ensure the platform—schedulers, runners, CI/CD, secrets, and monitoring—operates securely and reliably.

Formación

• Built and run AI agents in production with real users and failures.
• Production code in Python and TypeScript; fluent with LLM APIs and tooling.
• Hands-on infrastructure administration: schedulers, cloud services, CI/CD, secrets, DBs and monitoring.
• Security-minded with a practical approach.
• Strong documentation habits: specs, runbooks and change logs.

Responsabilidades

• Own all agents: health checks, logs, alerts, fixes and reviews.
• Create and test new agents from team needs and specs.
• Oversee the platform: schedulers, runners, CI/CD, secrets, monitoring.
• Enforce production standards: testing, logging and rollback for agents.
• Maintain memory and security: access rules, inventory, audit trails.
• Build integrations across CRM, email, apps, websites and analytics.
• Drive a unified customer journey across marketing, sales and product.

Conocimientos

AI agents

Python

TypeScript

LLM APIs

Security

Documentation

Communication

Operator mindset

Herramientas

Claude Code

Claude Agent SDK

Vector databases

Azure

Google Workspace admin

Monday.com

CRM APIs

Accounting APIs

Descripción del empleo

Location: Salt Lake City / Cottonwood Heights, Utah

Compensation: Competitive base salary plus performance bonus; equity available for the right person

The Opportunity

Our client is a small, fast-growing wellness technology company that already runs much of its business on AI agents. In the last few months, the founder and a small engineering team built a working AI operating system: an AI COO that tracks commitments across the company, a founder's assistant, a sales-call coach, prospecting pipelines, a shared memory layer and a live company dashboard. It runs on Claude Code and the Claude Agent SDK.

It grew fast and now needs an owner. You'll make it reliable, observable and secure, then build the next layer: one system that tracks the full customer journey across marketing, sales and product, with agents that act on it.

What You Will Own

• Every agent, running: health checks, run logs and alerts, same-day fixes, and regular reviews of whether each agent is actually useful to the person who owns it.
• New agents: turn a team's need into a spec, build it, test it on real work and hand it off.
• The platform: schedulers, runners, cloud routines, CI/CD pipelines, secrets management, monitoring and alerting.
• The production standard: every agent goes live through your review, with clear bars for testing, logging and rollback.
• Memory and security: company memory with strict access rules, least-privilege access for each agent, a full inventory of keys and connectors, and an audit trail.
• Integrations: the connections between CRM, email, the app, website, ad platforms, accounting and workspace tools, built alongside the product engineers.
• The growth engine: one tracked customer journey, from first touch to repeat order, that every team works from.

What You Bring

• You've built and run AI agents or automated workflows in production, with real users, real failures and real fixes.
• Production code in Python and TypeScript; fluency with LLM APIs, tool use and MCP; daily use of AI coding tools.
• Hands-on infrastructure administration: schedulers, cloud services, CI/CD, secrets, databases and monitoring.
• A serious but practical approach to security.
• Strong documentation habits: specs, runbooks and change logs someone else can follow.
• Calm under pressure, and plain, candid communication with leadership.

Nice to have: Claude Code or the Claude Agent SDK; vector databases, retrieval and evals; Azure, Google Workspace admin, Monday.com, CRM or accounting APIs; health-tech or privacy-by-design experience; having been a first technical hire.

This is not a prompt-writing or project management role. It's for an engineer who treats an agent as software with a job description: inputs, schedules, permissions, failure modes and a human owner who has to trust it.

The Environment

A lean, founder-led team where a small group does the work of a much larger one. You report directly to the founder and have your own budget for infrastructure and tooling. The role can grow into Head of AI Systems with a team as the company scales.

The role is fully on-site in the Salt Lake City area. Candidates who are willing to relocate on their own are welcome; relocation assistance and visa sponsorship are not available.