Applied AI Engineer

Build AI agents that help run real mines and processing plants: agents that query live sensor data, call physics-based models and flag problems before people notice them!

About IntelliSense.io

At IntelliSense.io, we are pioneering the future of industrial decision intelligence. Our platform is trusted by Tier 1 and Tier 2 global mining leaders and cutting-edge secondary metal recycling operations.

We combine digital twins of individual industrial processes with AI agents for machines and humans, including role-specific solutions like a "Digital Metallurgist" or "Digital Geologist", predictive ambient agents, and a 24/7 site memory that survives shift changes and staff turnover. As an AI-first company, we harness intelligence across everything we do to cut time-to-value and scale rapidly.

The role and why it matters

You will build the tools and LangGraph workflows that let agents query live plant data, call physics-based models, search documentation and show results as live charts and tables in Next.js. You will also help build ambient agents that monitor industrial sites 24/7, with tracing and evaluation so they are safe to rely on.

You bring modern Python (async, type hints), hands-on experience building an LLM agent or tool-calling app, SQL and RAG, good engineering habits, and curiosity beyond Python. Mentorship from senior engineers; mining knowledge not required.

What you'll do

• Build the tools our agents use: querying time-series, relational, graph and vector data (Postgres, MongoDB, Neo4j, Qdrant), RAG over documentation, and integrations with services such as Slack, Notion and ClickUp.

• Design and improve multi-agent workflows in LangGraph, including prompts, few-shot examples and agent configuration.

• Connect agents to our domain models for grinding, flotation and thickening circuits.

• Help build generative UI in Next.js with the Vercel AI SDK, so agents can show live charts, tables and widgets.

• Contribute to ambient agents that monitor the health of industrial sites and of our platform.

• Add tracing, evaluation and monitoring so agent reasoning is transparent, reliable and safe for industrial use.

• Help colleagues across the company use and improve our agents.

What you'll bring

You will ideally have a Master’s preferred in Computer Science, AI, Physics, Mathematics, or a related field.

Your technical skills should include:

• Modern Python: Strong foundation in Python (3.12+), specifically async/await and type hinting.

• Agentic Mindset: You’ve moved beyond simple prompts and are excited by the challenge of state management, tool-calling, and reliability in LLM systems.

• Full-Stack Curiosity: You are happy to dive deep into a Python backend but also open to investigating new APIs and data formats, shipping a React component, or tweaking a Dockerfile.

• First-Principles Thinking: You enjoy breaking down complex problems and are comfortable working in a field where the best practices change every week.

• Signal vs. Noise: You possess the critical thinking to separate first-principles engineering from marketing hype.

• Operational Rigor: An interest in building production-grade software with clean code, unit tests, and thoughtful documentation.

What we offer

• Unlimited holidays: we trust you to manage your time and your energy.

• True flexibility: work in a way that brings out your best results.

• Direct mentorship from senior engineers defining the architecture of industrial agentic systems.
• A diverse team spanning the UK, Ireland, Spain, Chile, Saudi Arabia, Australia and beyond.

Location

Cambridge, hybrid. 

Our process

A short call with our People team, a technical interview about a project you have built, a take-home exercise (3-4 hours) with a review session, a conversation with the team, and references. All stages can be held online. We aim to complete the process within 8 weeks.

Inclusion

We want our teams to reflect the range of people who use and are affected by our technology. We welcome applications from everyone and will make reasonable adjustments at any stage: just tell us what you need. If you meet most, but not all, of the essentials, we would still like to hear from you.