Software Engineer
Summary
Engineer at an early-stage, VC-backed AI startup in Melbourne building a real-time agent that joins live calls, ingests video/screen-share, and guides customers through software via natural conversation. Work spans LLM latency optimization, human-feeling audio/video handling, agent knowledge, and WebRTC call infrastructure at scale.
Salary: $200,000-$220,000 + super + (significant) ESOP
Looking to dive into something as fun as it is technically challenging?
This incredible team have built a real-time agent that joins a live call, ingests a continuous video feed and screen share, and holds a natural conversation while it guides customers through software. They have been building this for 18 months, have hit massive revenue milestones, are years ahead of the market and have a serious technical moat around their product.
It's already in production with Google, Canva and Apollo, and we reached $1M ARR in 8 months. They are backed by Bond, Felicis, Blackbird and Atlassian Ventures, and founded by an ex-Immutable / Atlassian team that has built and exited a company before. Enterprise demand is outrunning what they can currently serve.
The problems you'll work on:
- Sub-2-second conversations. From the moment a user stops speaking to the agent replying, across the whole pipeline. Context window management, model performance profiles and infra all have to line up, and every millisecond counts.
- Audio and video that feel human. Turn-taking, barge-in, echo remediation, transcript confidence, voice activity detection, and screen activity detection: knowing when to speak based on what's happening on someone's screen.
- Agent knowledge. Becoming a subject-matter expert in a customer's product before the call, recalling it fast enough to stay inside the latency budget, and keeping it current as the product changes.
- Scale. WebRTC call infrastructure for enterprise customers with variable load, run by a very small team.
- Deploying agents across whole organisations. The next frontier: how do you roll this out at scale?
How they work
Apart from their founders, they don't have assigned product managers. You are one ☺. With their advanced use of LLMs, this means you can talk to enterprise customers, form opinions on what a good product is, take an idea, make the case for it, prove it with a POC, spot the pattern across customers, agree the solution with the other engineers and then build it.
Apply if you
- Are a strong engineer with fundamentals you can reason about out loud, typically 4+ years in, ideally with time at a startup or scale-up
- Care about why you're building something and who it's for, not only how
- Have built things outside your day job: a side project, a startup, even one that didn't work out
- Follow what's happening in AI closely and can say what you've tried recently
- Can explain a hard technical trade-off to a customer or a non-technical teammate
- Want a big ambition for your career and see a company like this as a way to get there faster
Ready for your career-defining role? Apply now ! Can't wait to chat with you!