AI Engineer

Kimia is building the operating system for the $5T chemical enterprise - and we're just getting started.

The chemical industry runs on products that are technically complex, surrounded by tens of thousands of documents, and fragmented across silos like CRM, ERP, PIM, and regulatory systems. For decades, chemical companies have been managing catalogues of highly technical products with tooling built for commodities. The mismatch is staggering: a single specialty polymer might have hundreds of performance parameters, regulatory constraints across dozens of jurisdictions, and application data buried in PDFs from the 1990s - and the "system of record" is often a spreadsheet.

AI finally makes it possible to work with the complexity and nuance of this data. We're using that to build the data layer, workflows, and intelligence chemical enterprises should have had all along.

We're the first company since Canva to be funded at early stage by all three of Australia's top VC funds - which tells you something about the calibre of people who've looked closely at what we're building and said yes. Our customers include some of the largest chemical distributors and manufacturers in the world. The problems are real, the scale is real, and the work matters.

If you want to build something that genuinely changes how a trillion-dollar industry operates - this is the place.

The role

We're hiring an AI Engineer to work directly with founders and a small team of engineers. You'll own real product work end-to-end - not research experiments that never ship, but agentic systems that run in production and do real work for real customers.

This is one of the most interesting applied AI problems out there. Chemical data is genuinely hard: sparse, unstructured, technically dense, and full of edge cases where "close enough" isn't good enough. Retrieval alone doesn't cut it - the system needs to reason about what it's retrieved, handle ambiguity, and know when it's wrong. If you've been looking for a place where AI engineering is the product, not a feature bolted onto one, this is it.

What you'll do

- Design and ship agentic workflows, retrieval systems, and LLM features that run in production

- Build the evals, tracing, and guardrails that make AI features reliable enough for enterprise customers

- Work across the stack when needed - you'll spend most of your time in Python, but you'll touch the TypeScript codebase too

- Turn messy, ambiguous requirements (sometimes literally chemical spec sheets) into clear, buildable systems

- Keep the bar high on code quality, testing, and reliability - we're building software enterprises will run their business on

- Drive timelines and keep things moving, including pushing back when scope is wrong

The right candidate

This role is perfect for someone who's been building with LLMs hands-on, is energized by green-field problems, and cares as much about whether the system actually works in production as whether the demo looks good.

- You've shipped LLM features to real users and have scars to show for it - you know the difference between a prompt that works on five examples and one that works on five thousand

- You take evals seriously and have opinions about how to measure AI systems

- You take ownership and push for excellence - you're the kind of person who notices the thing that's broken and fixes it, even when it's not "your" code

- You don't wait to be told what to do

- You're comfortable figuring things out with limited resources and imperfect information

- You adapt fast when priorities change (and they will, especially in AI)

- You're always improving your craft - you read papers, you tinker, you have opinions about tools and models

Tech stack

Python, LLM tooling, agent frameworks, vector stores, eval pipelines. TypeScript, NestJS / Express, Next.js, React, PostgreSQL for the product surfaces you'll integrate with. Bonus points for AWS and CI/CD (GitHub Actions, CDK). We care less about which specific tools you've used and more about how quickly you can get productive in a new one - the AI stack changes every six months anyway.

What's unique about this opportunity

- Builder culture. You'll have high agency and make product decisions constantly. In the age of AI, taste and judgment are the most valuable skills an engineer can have - we want people who already think this way and want to sharpen it further.

- Competitive pay. Benchmarked to scale-up comp, plus a meaningful equity stake. You're an owner.

- On the technical frontier. We're figuring out what AI-first software looks like for an industry that's never had modern tooling. The problems are genuinely hard: the UI/UX, data models, and agent capabilities for this space don't exist yet. You're not re-implementing something that already works elsewhere - you're inventing it.

- Remote-first, AEST-anchored. We're distributed, but our core collaboration hours are Sydney time. We'll host you in Sydney for onboarding so you can meet the team in person.

- No-nonsense recruitment. Informal chats with the founders and a technical conversation with the team. No LeetCode, no take-home busywork - we want to see how you actually think and build.