Applied AI Engineer (AI)
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Описание вакансии
TL;DR
Applied AI Engineer (AI): Building LLM-powered features to automate complex real-world workflows in industries like aerospace and healthcare with an accent on model behavior and reliable production deployment. Focus on designing multi-step agents, building robust evaluation frameworks, and extracting structured data from complex documents.
Location: On-site in London, UK (Visa sponsorship and relocation available)
Company
is a venture-backed startup building systems to accelerate manual workflows in physical-world industries like aerospace, healthcare, and energy.
What you will do
- Deploy AI features to production from day one and own user-facing surfaces.
- Manage the end-to-end lifecycle from problem definition and prompting to evals and monitoring.
- Design multi-step agents and orchestration for complex domain workflows.
- Extract structured data from messy real-world documents such as PDFs, scans, and forms.
- Collaborate with customers to debug outputs and iterate on system performance.
Requirements
- Strong proficiency in Python and modern LLM tooling (OpenAI, Anthropic, Azure OpenAI SDKs).
- Experience building production AI systems beyond simple API wrappers.
- Expertise in prompt engineering, evals, and analyzing model failure modes.
- Ability to work across the stack using TypeScript, React, and Python.
- Must be able to work on-site in London.
Nice to have
- Founder or ex-founder experience.
- Background in early-stage or stealth startups.
- Experience with founder fellowships such as EF, YC, or Antler.
Culture & Benefits
- High-ownership environment with a tight-knit, world-class team.
- Visa sponsorship and relocation costs fully covered.
- Daily meals and tools provided.
- Fast-paced culture with direct impact and no corporate politics.
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 →