About the role
About the Role
Build and maintain an AI agent that reviews customer code for OpenTelemetry instrumentation issues, proposes fixes, and supports migrations away from proprietary agents. You will work across a TypeScript control plane, sandboxed agent runs, and the agent harness, contributing to both product features and open source tools.
What You'll Do
* Build agent features from specification through production, including GitLab support and workflows that turn telemetry findings into code fixes across customer repositories.
* Improve agent reliability with structured outputs, retries, long-running sandbox jobs, and consistent findings across runs.
* Create evaluations to measure the quality and cost of prompt and model changes.
* Develop context that helps agents safely review and modify Go, Java, Kotlin, TypeScript, Python, and .NET code.
* Maintain open source agent skills, keep them current with OpenTelemetry releases, review contributions, and use evaluations to assess changes.
* Partner with customer engineers on deliveries alongside the Customer Success team.
What We're Looking For
* At least 5 years of software engineering experience and strong TypeScript proficiency.
* Experience shipping and maintaining an LLM agent or AI-assisted product in production.
* Hands-on experience measuring LLM quality with evaluations, golden sets, or production signals.
* Ability to read and safely modify Go, Java, Kotlin, Python, and .NET code.
* Experience with OpenTelemetry instrumentation, APIs, webhooks, durable workflows, and Postgres.
* Clear async communication skills, including design documents, pull request descriptions, and written summaries.
* Experience with agent skills, plugins, or MCP servers; open source maintenance, GitHub or GitLab integrations, sandboxed code execution, Cloudflare Workers, or Bun is useful.
Location
Remote, based in Argentina. Applicants must be based in Europe or Latin America. Visa sponsorship is not available.