AI Automation Engineer (Claude Code)

Applaudo Studios

2d

About The Position

Applaudo is hiring an AI Automation Engineer to adapt an existing Claude Code Jira support-triage routine, originally built for another product, for a channel incentive platform consisting of about 70 Node.js and .NET microservices. The engineer will own the work from discovery through handover, partner with a platform-team subject matter expert, and present design alternatives to the client’s architect for a decision. The role calls for a persistent, self-directed person who can iteratively refine automation until its output is reliable.

The work includes investigating the reference routine, support queue, repositories, and relevant code paths; proposing the adapted design and its safety rules; building and scheduling the routine in a cloud environment; testing it on tickets and reviewing trial comments; maintaining the prompt in source control; and training platform engineers. The engineer will also document maintenance procedures and hand ownership to a named platform-team owner. The position operates under a contractor framework for an initial period of three months and requires availability to work on U.S. Eastern Time, starting at 8:00 AM.

Requirements

• Bachelor’s degree in Computer Science, Software Engineering, Computer Engineering, or a related field is desired, or equivalent professional experience.
• At least 5 years of professional software development experience, including production C# and .NET code and the ability to read older .NET Framework code.
• At least 1 year of hands-on experience building LLM agents or automations.
• Proven delivery of Claude skills or Claude Code automations used by a real team.
• Experience designing agents through a written procedure, including hard rules, investigation steps, classification logic, and output formats.
• Experience applying least-privilege design to agents, with limits enforced in the environment and permissions.
• Experience running scheduled or unattended agents through Claude Code routines or an equivalent scheduler, including cloud environments and connectors or MCP servers.
• Hands-on experience with Jira Service Management, including JQL, saved filters, boards, the service desk REST API, and the distinction between internal and customer-visible comments.
• Experience with GitHub workflows, including apps, branches, draft pull requests, and code review.
• Ability to read and write production C# and .NET code and read older .NET Framework code.
• Ability to read Node.js code well enough to trace a feature across microservices.
• SQL skills to read a schema and write scripts that use only existing tables and columns.
• Experience improving prompts based on failures by reading session transcripts, identifying the root cause of incorrect output, and changing the procedure.
• Ability to learn a large codebase and support process independently, with limited time from subject matter experts.
• Sound judgment about customer-facing risk, defaulting to internal communication when a public action cannot be undone.
• Comfort proposing options with clear trade-offs and accepting the architect’s final decision.
• Ability to teach developers who use Claude but have never built a routine.
• Advanced English proficiency.
• Strong verbal and written communication skills, including technical discussions, stakeholder interactions, and documentation.
• Availability to work on U.S. Eastern Time, starting at 8:00 AM.

Nice To Haves

• Experience with agent observability, evaluation, or reliability tools.
• MCP server development experience.
• Experience with support operations or IT service management processes.
• Node.js development experience.
• Experience with channel incentive, rebate, or other financial-flow platforms.
• Anthropic, AWS, or other AI certifications.

Responsibilities

• Conduct a one-week discovery of the reference routine and its documented gotchas, the support queue and its configuration, and the repositories and code paths the routine must read.
• Propose an adapted design covering JQL scope, classification rules, investigation depth, SQL and schema rules, internal and public comment rules, and permissions; present alternatives to the client’s approver for a decision.
• Build the routine by writing its prompt as a numbered procedure with hard rules, configuring the cloud environment, connectors, and schedule, and keeping the prompt in source control.
• Test the routine by hand on one or two tickets, review every comment it posts during the trial, and adjust the prompt for each class of mistake while keeping a dated record of every change.
• Open draft pull requests only through the routine, write C# code when needed, and follow the team’s standard review and test process.
• Train platform engineers on how the routine works, what its labels mean, how to read a session transcript, and how to edit the prompt safely.
• Write a maintenance guide and hand over ownership of the routine to a named owner on the platform team.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior