Principal AI Engineer

Bluebird; Greater London; Remote; GBP 120,000 - 200,000; Full time

Job summary

Bluebird is building an AI-native learning platform and seeking a senior AI engineer to own core systems for an AI Tutor and an AI Content Creator. You’ll ship production-grade agentic workflows, optimize costs, latency and reliability, and mentor a small team as the product scales.

You will design end-to-end architectures, work with LangGraph/LangChain and vector stores, and drive quality controls to ensure the product actually teaches and delivers real user value at scale.

Qualifications

• Shipping AI systems to real users in production.
• Experience at principal or staff level on complex production AI systems.
• Hands-on experience with model and cost optimization.
• Genuine interest in education or EdTech/AI-education product experience.

Responsibilities

• Design and ship agentic workflows in Python with LangGraph or LangChain.
• Own systems end to end, from architecture through production, monitoring and iteration.
• Optimize for cost, latency and reliability at scale.
• Mentor a small team and shape zero-to-one product direction.

Skills

Shipping AI systems

Principal/staff level experience

Model and cost optimization

Education/EdTech interest

Education

Startup experience (YC)

Tools

LangChain

LangGraph

Open-weight models

Job description

Salary: up to 200,000 GBP base, plus equity and bonus

About the role

We are partnering with a category-leading online learning company that is rebuilding its product around AI-native, personalized tutoring. Following the acquisition of an AI-native learning startup, they are assembling a small, high-calibre AI engineering team to build two core systems from the ground up: an AI Tutor that delivers real-time, adaptive teaching (finally cracking one-to-one tutoring at scale), and an AI Creator that turns expert knowledge into high-quality, interactive learning content. The early AI product is already outperforming their legacy platform on learner satisfaction and conversion, and you would have real ownership as the team scales.

What you will build

• The core agentic systems behind a real-time AI tutor and an AI content generator.
• Production LLM systems that are fast, reliable and cost-efficient, including the move from frontier models toward open-weight models without losing quality.
• The evaluation frameworks and human-in-the-loop quality controls that prove the product actually teaches.

What you will do

• Design and ship agentic workflows in Python with LangGraph or LangChain, prompts and vector databases.
• Own systems end to end, from architecture through production, monitoring and iteration.
• Optimize for cost, latency and reliability at scale.
• Mentor a small team and shape zero-to-one product direction.

What we are looking for

• A builder first: strong software-engineering foundations and a track record of shipping agentic AI systems to real users in production.
• Proven ability to operate at principal or staff level on complex production AI systems.
• Hands-on experience with model and cost optimization, and with evaluating output quality (golden sets, LLM-as-judge, user feedback).
• Genuine interest in education, or prior EdTech or AI-education product experience (especially for the Creator work).
• Background at a fast-moving startup (YC or similar) or as a founder.
• Experience fine-tuning or adapting open-weight models.
• Work on educational content quality or data-labeling pipelines.

Logistics

Remote-friendly with occasional travel for collaboration. Permanent role, with visa sponsorship and relocation available for the right person.