Senior AI Engineer
New
T
TenaraiAI engineering
Workplace type: remote; Country code: PLFull-TimeSenior
Salary170 - 200 PLN per hour
Apply NowOpens the employer's application page
Job Details
- Languages
- En B2
- Required Skills
- Node.jsArtificial IntelligenceJenkinsTypeScriptGitHub ActionsPlaywright
Requirements
- Have built agent extensions used by other people, such as skills, instructions, hooks, MCP servers, custom tools, subagents, or plugins.
- Know an agentic coding harness such as Claude Code, Copilot CLI, Codex, or Cursor at the configuration and extension level.
- Explain how a harness loads instructions, fires hooks, handles permissions, and connects tools and MCP servers.
- Understand agent loops, context-window budgets, compaction, prompt caching, delegation, and non-determinism, with examples from your own work.
- Have built an evaluation or regression gate for prompts, skills, or agent behavior and be able to explain what it checks and where it is useful.
- Measure workflow costs, such as tokens, model mix, or cache hits, and demonstrate a cost reduction without quality loss.
- Show experience splitting work across agents or worktrees, writing briefs, integrating results, and catching bad output before merge.
- Understand product development and release processes as well as quality engineering, including test strategy, CI pipelines, failure triage, or flakiness fixes.
- Be able to describe safeguards for personal, payment, or customer data, including data from test environments.
- Demonstrate experience enabling adoption by capturing expert knowledge, using working demos and measured results, coaching teams, and embedding practices in tooling.
- TypeScript / Node.js, Playwright, and REST/XML/JSON API testing.
- GitHub PRs, Actions and reusable workflows; Jenkins; Allure.
Responsibilities
- Build and maintain skills, instructions, hooks, MCP and CLI tools, and subagent definitions for a plugin used by several agent clients.
- Design tools—including names, descriptions, parameter shapes, output sizes, and error messages—so agents select and use them correctly.
- Keep always-loaded context small and load procedures on demand; investigate and fix agent stalls, loops, and ignored instructions.
- Gate harness changes with deterministic evaluations and CI.
- Run work in parallel across agents and worktrees, select models for tasks, and make workflow costs visible.
- Work with QA and product engineers to capture tacit knowledge and turn it into reusable skills and decision records.
- Demonstrate agent performance on manual and exploratory testing tasks using before-and-after measurements.
- Build guardrails that keep personal, payment, and customer data out of prompts, logs, reports, and repositories.
View Full Description & ApplyYou'll be redirected to the employer's site