Вакансия из Hirify Global, списка международных tech-компаний Для мэтча и отклика нужен Plus
Пожаловаться
Зарплата и рынок
ЗП не указана
На похожих ролях по рынку
$83к$105кмедиана$149к
45 в выборке
На СНГ рынке: $4.3к/мес ($3.6к - $5.1к)
88
Крутая вакансия
развернуть
Чётко очерченная роль по развитию внутренней AI-платформы в растущем финтехе даёт заметную ответственность и работу с быстро развивающимися агентными инструментами, хотя требования к опыту очень специфичны.
Кликните для подробной информации
Высокий порог входаСовременный стекЧёткая зона ответственностиФинтех-направление
Оценка от Hirify AI
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
Agentic Tooling Engineer (AI): Building and maintaining shared infrastructure for an internal agentic framework with an accent on agent distribution, evaluation, certification, and deterministic platform checks. Focus on designing reliable tooling for non-deterministic systems, analyzing agent feedback loops, and driving adoption across engineering teams.
Location: Remote within the European Union
Company
hirify.global is a B2B payments platform providing commercial card issuing, payment accounts, FX, spend management, and embedded payment infrastructure for businesses, fintechs, software platforms, and banks.
What you will do
Build and maintain shared tooling for agent distribution and installation, shared skills and agents, linting, evaluation runners, and harness interfaces.
Support engineers building harnesses, skills, agents, and other agentic tooling in line with the framework requirements.
Build evaluations and certification processes to verify that harness changes improve outcomes before release.
Move stable, rule-based prompt steps into deterministic platform checks in collaboration with Developer Experience and Platform Infrastructure teams.
Collect user feedback, resolve underlying friction, add regression tests, and measure delivery outcomes while driving adoption across the technology organisation.
Improve the framework and help define new processes and agents on top of it.
Requirements
Deep, daily hands-on experience with a wide range of agentic tools, including Claude Code, Codex, Cursor, Windsurf, Aider, Devin, and CodeRabbit, plus the infrastructure on which agents run.
Experience building agentic tooling used by other engineers, such as skills, plugins, hooks, subagents, MCP servers, or custom agents.
Experience with LLM evaluation and context engineering, including evaluations for non-deterministic systems, failure-trace analysis, and understanding how context size, freshness, and consistency affect agent behaviour.
7+ years of software engineering experience on production systems, including developer tooling or an internal platform with real users. At Staff level, experience setting technical direction across several teams is required.
Strong system design and a verification mindset, including the ability to distinguish deterministic controls from controls that can remain with a model.
Strong Python, Git, and GitHub experience, including Actions, rulesets, and CODEOWNERS; ability to read Java and Kotlin with ease.
Precise technical writing and the ability to influence how other engineers work without direct authority.
Nice to have
Experience in a regulated environment such as payments or banking.
Security engineering background.
Experience maintaining an open-source project.
Culture & Benefits
Full-time remote work with flexibility within the European Union.
Growing engineering organisation with broad responsibilities, knowledge exchange, and a focus on excellence.
Flat hierarchy, transparent communication, and a relaxed professional atmosphere.
Choice of Windows or Mac as the preferred operating system.
Attractive remuneration, monthly mobility benefit, Wellhub membership, and a hirify.global Card with monthly credit.
Inclusive hiring process with accommodations available during interviews.
Будьте осторожны: если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого - это мошенники. Обязательно жмите "Пожаловаться" или пишите в поддержку. Подробнее в гайде →