The posting
Imagine having the power to stress-test an entire power grid against a hurricane or thunderstorm before the clouds even gather. That is the reality we are creating at Neara.
We use advanced machine learning to create engineering-grade, physics enabled digital twins of electricity grids across four continents, this helps asset owners understand their biggest challenges and bring the most viable solutions to life across millions of kilometres of infrastructure.
By simulating extreme weather and structural stress at a network-wide scale, we empower the world’s largest utilities to pinpoint risks, optimise investments and build a more resilient global energy future.
Our team is a collection of brilliant minds who are fanatical about making a tangible difference in the real world, utilising AI and machine learning to accelerate everything from data classification to complex scenario analysis. We have built a special culture where innovation thrives because everyone owns the mission and we need smart, creative people to help us scale this impact to every corner of the globe.
ABOUT THE ROLE
Utilities come to Neara with questions that used to take months of engineering work to answer. Which of our poles can safely stay in service? Where will vegetation contact the line in the next storm season? What happens to this circuit if we add 200MW of solar? Today, getting to those answers means knowing your way around a deep, powerful platform.
We are building a layer of AI-powered features on top of that platform so that far more people can get to those answers, far faster — and so that the answer comes back grounded, traceable, and defensible to an engineer who has to sign their name against it.
This is a new team. As one of our first AI Product Engineers you will be shipping LLM and agentic features into a product that utilities use to make decisions about critical infrastructure, and helping set how we build with AI across engineering: our evaluation discipline, our guardrails, and our judgement about what these systems should and should not be allowed to do.
To be clear about what this role is not: our Digitisation and ML Platform groups already build and run the computer vision and machine learning that turns lidar and imagery into a digital twin. You will work closely with them, but you are not doing that. Your job is the product layer — LLMs, agents, retrieval and tool use, grounded in the twin, the physics engine, and the network data our customers already trust.
WHAT YOU WILL DO
- Design, prototype and ship AI-powered features end to end — from a problem raised by a utility engineer through to something running in production and being used.
- Build agentic and retrieval-based systems that are grounded in the Neara digital twin, our simulation outputs and our internal expression language, so that every answer can be traced back to the model it came from.
- Own the hard part: making AI output trustworthy in a domain where a confidently wrong answer about clearance or pole loading is a safety problem, not a bad user experience. Design the guardrails, the refusals, the confidence signals and the human-in-the-loop checkpoints.
- Build the evaluation discipline behind all of it — eval sets, regression testing when a model or prompt changes, offline and online quality signals, observability, prompt and context versioning, and cost and latency budgets.
- Work alongside our power systems and GIS specialists to make sure what we ship is engineering-grade, not just plausible.
- Spend real time with customers. Work with Product, Design and our customer-facing teams to decide what we build, kill the ideas that don't hold up, and iterate on what does.
- Set the standard for how Neara engineers build with AI, including how we use coding agents in our own workflow.
WHAT WE ARE LOOKING FOR
Essential
- [5]+ years building and shipping production software, with real depth in backend or full-stack product engineering. We need strong software engineers first.
- You have shipped AI features to paying customers in a commercial product, and owned them after launch. Not a prototype, a hackathon or a weekend project — something real users depended on, that you watched break and fixed. You can talk about who used it, how many, and why it mattered to them.
- Hands-on experience with the things that make LLM products actually work: retrieval, context engineering, tool use and agent orchestration, structured output, and prompt design.
- You take evaluation seriously and can describe an eval approach you built, what it caught, and what it missed.
- Sound judgement about where AI should and should not be used, and the instinct to design for failure modes rather than for the happy path.
- Strong product sense and comfort working directly with users & product teams — including expert users who will tell you bluntly when the software is wrong.
Nice to have
- Experience shipping into regulated or enterprise environments: data residency, auditability, private deployments, security review.
- Geospatial, 3D, CAD, simulation or other engineering software backgrounds.
- Experience with our stack: TypeScript/React, Python, Java, Dart/Rust - Experience in our exact stack is not required — we care more about how quickly you pick a new one up, and how deeply you already use AI coding tools in your day-to-day work.
- Utilities, energy or critical infrastructure domain knowledge.
WHAT WE OFFER
- Competitive salary and meaningful ESOP
- Fully flexible working arrangements with a well-stocked office in Redfern
- Regular office events
- The real benefit: working on a complex, innovative and industry-leading product that makes a real difference in the world around us.
To apply, please use the online application link below. Neara values diversity, belonging and equal employment opportunities. We encourage individuals from all backgrounds to apply.
Please note that you will require current and unrestricted working rights to be considered for the role. Unfortunately, we are not in a position to offer visa sponsorship at this time.
No agencies or third-party service providers, please.
#LI-AP1
