Вакансия из Hirify Global, списка международных tech-компаний Для мэтча и отклика нужен Plus
Пожаловаться
Зарплата и рынок
ЗП не указанаразвернуть
На похожих ролях по рынку
$82к$105кмедиана$125к
43 в выборке
На СНГ рынке: $4.3к/мес ($3.6к - $5к)
84
Хорошая вакансия
развернуть
Понятная зона ответственности за production RAG и качество поиска в быстрорастущем AI-стартапе в legal tech, но смена приоритетов и длительный процесс из пяти этапов добавляют неопределённости.
Нестабильные приоритеты
Пять этапов отбора
Актуальный стек
Чёткая ответственность
AI в legal tech
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
Senior RAG Engineer (AI): Building and operating retrieval pipelines for confidential legal documents with an accent on parsing, embeddings, hybrid search, reranking, and multilingual retrieval quality. Focus on designing agentic retrieval loops, evaluating production regressions with measurable metrics, and keeping secure, tenant-isolated services reliable under load.
Location: Fully remote across the EU/EEA, including Norway; listed locations include Oslo, Wrocław, and Chicago. The role requires existing authorization to work in the applicable country, and visa sponsorship is not available.
Company
hirify.global is a fast-growing legal tech and agentic AI company building products at the intersection of law, technology, and intelligence.
What you will do
Own the end-to-end retrieval pipeline, including document parsing, chunking, embeddings, indexing, ingestion, and reindexing.
Design hybrid search with semantic and keyword retrieval in Qdrant, metadata filtering, ranked-list fusion, and reranking.
Build agentic retrieval capabilities such as query decomposition, document-navigation tools, multi-step loops, and cost and latency controls.
Develop evaluation infrastructure with golden sets, retrieval metrics, regression tests, and tracing from incorrect answers to source chunks.
Ship dependable Python and FastAPI services with ingestion, embedding, and reindexing background jobs that operate under load.
Maintain strict tenant isolation and data security while improving retrieval across client languages, including multilingual morphology and inflection challenges.
Requirements
At least five years of production backend experience, including at least two years shipping retrieval or RAG systems used by real users.
Experience diagnosing and fixing production retrieval regressions, explaining the cause, investigation, and before-and-after metrics.
Experience building and maintaining golden sets, including query volume, labeling process, trusted metrics, and decisions driven by evaluation results.
Experience running vector indexes in production, selecting index parameters, managing memory and latency trade-offs, and reindexing without downtime.
Strong Python and practical experience with FastAPI or a close equivalent, PostgreSQL/SQL, and Redis under load.
At least five years of Python experience and at least two years of FastAPI experience.
At least two years of production experience with vector databases; the stack uses Qdrant.
Experience with embedding models, retrieval-quality metrics, OCR-heavy document ingestion at scale, BM25, NDCG, recall@k, and search abstractions spanning more than one backend.
Ownership of systems end to end, including maintenance and unassigned technical debt, plus comfort working in a startup where decisions and priorities can change.
Existing right to work in the applicable country is required; employer visa sponsorship is unavailable.
Nice to have
Side projects, benchmarks, or repositories demonstrating independent exploration of vector databases or LLMs.
Experience with AI coding assistants and a considered approach to reviewing AI-generated code.
Culture & Benefits
Fully remote work across the EEA, including Norway, with employment through a local entity or employer of record where applicable.
Flexible, English-speaking environment with collaborative peers.
Opportunity to influence product and culture while working in an AI venture.
Fast-changing startup environment requiring comfort with ambiguity, ownership, and proactive problem-solving.
Hiring process
Five stages: online coding challenge, screening call, technical home assessment, whiteboard challenge with presentation of the assessment, and CEO interview.
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 →