Zello Applied AI Engineer at Zello responsible for building AI agents from prototype to production and managing their ongoing health. This includes monitoring quality, managing human reinforcement workflows, and driving continuous improvement.
Responsibilities
Build AI agents and automations end-to-end: from scoping the use case through deployment and ongoing maintenance
Write production Python code that integrates LLM APIs (prompt construction, response handling, context management, tool use) into real workflows
Connect AI tools with Zello's systems (Slack, Jira, HubSpot, Snowflake) through APIs, handling authentication, rate limits, error cases, and logging
Monitor deployed agents in production: track quality metrics, triage failures, and ship improvements based on real usage data
Manage human reinforcement operations: review agent outputs, maintain feedback loops, and tune agent behavior based on reinforcement signals
Build and maintain evaluation harnesses that catch regressions and measure agent quality programmatically
Create reusable components, patterns, and documentation that raise the bar for future development on the team
Communicate clearly with technical and non-technical stakeholders about what you've built, what's working, and where things need attention
Qualification
You pick up new frameworksYour code is clean and documentedThis role is notA research roleA data engineering roleA DevOps or infrastructure roleA solo project
Required
You've written production Python and can point to real things you've built with it: tools, integrations, automations, shipped products. Not just notebooks or coursework.
You understand LLM APIs at a practical level. You can construct prompts, manage context windows, reason about token economics, and work with tool-use patterns.
You decompose messy problems into clean components with well-defined interfaces. When you describe a system you've built, people can follow the logic because you think in terms of abstractions, dependencies, and failure modes.
You've integrated systems via APIs before. You can read API docs, handle auth, manage rate limits, and deal with the inevitable edge cases of real-world integrations without getting stuck.
You're comfortable with operational ownership. You don't treat deployment as the finish line. You monitor what you build, notice when things drift, review agent outputs, and do the sometimes unglamorous work of keeping AI systems healthy in production.
You pick up new frameworks, APIs, and domains quickly. You can point to examples of going from zero to productive in an unfamiliar area.
Your code is clean and documented. Other people can read it, understand it, and extend it without needing a walkthrough from you.
This role is not
A research role. We're building on top of foundation model APIs, not training models or publishing papers.
A data engineering role. The existing team covers data infrastructure. You'll consume data, not build pipelines.
A DevOps or infrastructure role. You'll deploy your own agents, but you won't be managing servers or building CI/CD from scratch.
A solo project. You'll work closely with the Data & AI team and cross-functional stakeholders who use what you build.
