Job Description
Howdy, we’re Heidi
“The AI startup growing faster than Canva”
That’s what the Financial Review called us. In 18 months, we supported over 73 million patient visits and become one of the fastest-growing companies in the world.
We pivoted from broad healthcare AI to building Earth’s finest AI Care Partner. Today, we support over 2 million patient sessions weekly across 116 countries and over 110 languages. Hundreds of thousands of clinicians use Heidi to complete documentation.
Our mission is simple: strengthen the human connection at the heart of healthcare.
We’ve found product-market fit with individual clinicians through our freemium medical scribe, transforming unstructured clinical visits into structured text artefacts. Clinicians and organizations quite like it. Now, we embark upon consuming more than just documentation. Every new job a clinician delegates to Heidi makes patients feel more attended to, cleans up health system logjams, and lets clinicians be clinicians again.
That’s where you come in.
The role
We’re looking for a Product Manager to own the AI models that power everything Heidi does. Someone who thinks the best platform teams exist to make product teams faster.
You will own Heidi’s models platform: evaluation pipelines, fine-tuning infrastructure, model routing, and safety systems. Hundreds of thousands of clinicians across 116 countries use these models in clinical settings every week. You’ll work with engineers and researchers, partner with product PMs and clinical safety, and stay close enough to product teams that you know what they need before they file a ticket.
Every user-facing product at Heidi depends on what your team builds. When a clinician gets an accurate note, when a referral letter captures the right history, when a model update ships without breaking three downstream products: that’s your work.
This role will be based in either our Sydney or Melbourne office.
We don’t care about logos; the traditional insignia of competence. We’ll evaluate senior well-credentialed candidates and young, hungry hopefuls alike.
What you’ll do
Own product strategy and roadmap for Heidi’s models platform (evaluation, safety, model routing, fine-tuning infrastructure), setting clear goals and being held accountable to achieving them
Prioritise your team’s work across enablement requests, model safety and quality, and bets on new capabilities
Figure out where product teams get stuck on your models and fix the platform so they don’t
Build eval tooling and fine-tuning workflows that your engineers and product teams can actually use in clinical settings
Decide what to improve next by reading clinician feedback, model quality signals, and what product teams are asking for
Allocate engineering capacity across product teams who all want more than you can give, and tell them clearly what you’re deferring
Work with your engineers on eval design, fine-tuning trade-offs, and model architecture decisions at a technical level
Set model quality and safety targets grounded in clinical outcomes (did the note capture the right diagnosis? did the referral letter contain the right history?)
Spot infrastructure that two product teams are building separately and consolidate it
Watch foundation model developments and decide when to rip up your roadmap
If we’d worked together the last 6 weeks, you’d have:
Defined an evaluation framework for model quality that your engineers actually use
Made a clear ship/hold decision on a model update under pressure from a product team and communicated the rationale to leadership
Identified overlapping model capability requests across two product teams and proposed shared infrastructure
Built a 90-day roadmap that balances enablement requests with your own priorities for model quality
Had a productive disagreement with a senior engineer about prioritisation and reached a resolution you both committed to
What you’ll need
BA/BS in a technical or analytical field (e.g. Computer Science, Engineering, Information Systems, Analytics, Mathematics, Physics, Applied Sciences, or related)
4+ years working on AI platform, infrastructure, or model-adjacent products, though we care more about what you’ve built than time served
A romantic streak about software and a belief that great design transforms someone’s day
Technical depth on model evaluation, fine-tuning, and production AI systems; you’ve designed eval suites and debugged model regressions
Genuine curiosity about what models get wrong in clinical settings and why
Technical enough to hold your own with engineers, credible enough to present safety trade-offs to leadership
You build with AI tools (Cursor, Claude Code, whatever ships faster) and can show what you’ve made with them
Fluency with core LLM concepts and systems (prompting, fine-tuning, embeddings, retrieval, evaluation) and the judgment to translate these into reliable, user-facing products
You default to building over requesting
Willingness to update your views when the technology shifts, which it does roughly quarterly
Strong opinions, weakly held; you’ll shift the room when you’re right
Data fluency with diagnostic teeth; you can read evaluation results and distinguish a real regression from noise
If you answer ‘NO’ to these questions, this may not be the job for you:
Are you an execution powerhouse?
Have you shipped something this week?
Are you good at games?
Did you have a weird teenage hobby?
Are you able to execute without a legion of data analysts, product marketers, and research coordinators at your beck and call?
Does the prospect of making health systems a lot nicer make you feel fuzzy inside?
Why Heidi?
Our health systems hurt everybody. Systems are log-jammed, clinicians’ lives degrade in quality, and patients churn through the meat-grinder.
This is neither a special nor interesting insight.
However, it means we’re solving a real problem. Hundreds of thousands of clinicians use Heidi. Every clinician using Heidi gets hours back in their day. This is the starting point to build the product of your dreams: immense distribution, and a world of potential.
You’ll join a team that doesn’t celebrate fundraise milestones, imaginary valuations, and glossy PR. We live and breathe the pain of modern health systems, and won’t rest until we’ve exacted the change we’d like to see. We’re medicos, engineers, and designers who’ve felt (on every side of the equation) what non-care feels like - the moral and practical toll as a provider or receiver.
The bad news is it’s not an easy job - you will tear your hair out more than once. Modern technology company concepts don’t transpose unto healthcare easily. The lows are really low.
The good news is you will trust and admire everybody you work with, and rest easy knowing you’re doing THE defining work of your career. The highs are blindingly high.
Heidi will take care of you. We offer a generous $500 annual development budget and serious equity.
True A-players progress extremely fast here. The nature of the scale-up game is demanding, but we don’t track hours or micromanage - you’re trusted to perform. We value mental health and adapt to your schedule. We operate on outcomes > inputs, not process theatre. We all take the bins out, metaphorically and literally.
“The AI startup growing faster than Canva”
That’s what the Financial Review called us. In 18 months, we supported over 73 million patient visits and become one of the fastest-growing companies in the world.
We pivoted from broad healthcare AI to building Earth’s finest AI Care Partner. Today, we support over 2 million patient sessions weekly across 116 countries and over 110 languages. Hundreds of thousands of clinicians use Heidi to complete documentation.
Our mission is simple: strengthen the human connection at the heart of healthcare.
We’ve found product-market fit with individual clinicians through our freemium medical scribe, transforming unstructured clinical visits into structured text artefacts. Clinicians and organizations quite like it. Now, we embark upon consuming more than just documentation. Every new job a clinician delegates to Heidi makes patients feel more attended to, cleans up health system logjams, and lets clinicians be clinicians again.
That’s where you come in.
The role
We’re looking for a Product Manager to own the AI models that power everything Heidi does. Someone who thinks the best platform teams exist to make product teams faster.
You will own Heidi’s models platform: evaluation pipelines, fine-tuning infrastructure, model routing, and safety systems. Hundreds of thousands of clinicians across 116 countries use these models in clinical settings every week. You’ll work with engineers and researchers, partner with product PMs and clinical safety, and stay close enough to product teams that you know what they need before they file a ticket.
Every user-facing product at Heidi depends on what your team builds. When a clinician gets an accurate note, when a referral letter captures the right history, when a model update ships without breaking three downstream products: that’s your work.
This role will be based in either our Sydney or Melbourne office.
We don’t care about logos; the traditional insignia of competence. We’ll evaluate senior well-credentialed candidates and young, hungry hopefuls alike.
What you’ll do
Own product strategy and roadmap for Heidi’s models platform (evaluation, safety, model routing, fine-tuning infrastructure), setting clear goals and being held accountable to achieving them
Prioritise your team’s work across enablement requests, model safety and quality, and bets on new capabilities
Figure out where product teams get stuck on your models and fix the platform so they don’t
Build eval tooling and fine-tuning workflows that your engineers and product teams can actually use in clinical settings
Decide what to improve next by reading clinician feedback, model quality signals, and what product teams are asking for
Allocate engineering capacity across product teams who all want more than you can give, and tell them clearly what you’re deferring
Work with your engineers on eval design, fine-tuning trade-offs, and model architecture decisions at a technical level
Set model quality and safety targets grounded in clinical outcomes (did the note capture the right diagnosis? did the referral letter contain the right history?)
Spot infrastructure that two product teams are building separately and consolidate it
Watch foundation model developments and decide when to rip up your roadmap
If we’d worked together the last 6 weeks, you’d have:
Defined an evaluation framework for model quality that your engineers actually use
Made a clear ship/hold decision on a model update under pressure from a product team and communicated the rationale to leadership
Identified overlapping model capability requests across two product teams and proposed shared infrastructure
Built a 90-day roadmap that balances enablement requests with your own priorities for model quality
Had a productive disagreement with a senior engineer about prioritisation and reached a resolution you both committed to
What you’ll need
BA/BS in a technical or analytical field (e.g. Computer Science, Engineering, Information Systems, Analytics, Mathematics, Physics, Applied Sciences, or related)
4+ years working on AI platform, infrastructure, or model-adjacent products, though we care more about what you’ve built than time served
A romantic streak about software and a belief that great design transforms someone’s day
Technical depth on model evaluation, fine-tuning, and production AI systems; you’ve designed eval suites and debugged model regressions
Genuine curiosity about what models get wrong in clinical settings and why
Technical enough to hold your own with engineers, credible enough to present safety trade-offs to leadership
You build with AI tools (Cursor, Claude Code, whatever ships faster) and can show what you’ve made with them
Fluency with core LLM concepts and systems (prompting, fine-tuning, embeddings, retrieval, evaluation) and the judgment to translate these into reliable, user-facing products
You default to building over requesting
Willingness to update your views when the technology shifts, which it does roughly quarterly
Strong opinions, weakly held; you’ll shift the room when you’re right
Data fluency with diagnostic teeth; you can read evaluation results and distinguish a real regression from noise
If you answer ‘NO’ to these questions, this may not be the job for you:
Are you an execution powerhouse?
Have you shipped something this week?
Are you good at games?
Did you have a weird teenage hobby?
Are you able to execute without a legion of data analysts, product marketers, and research coordinators at your beck and call?
Does the prospect of making health systems a lot nicer make you feel fuzzy inside?
Why Heidi?
Our health systems hurt everybody. Systems are log-jammed, clinicians’ lives degrade in quality, and patients churn through the meat-grinder.
This is neither a special nor interesting insight.
However, it means we’re solving a real problem. Hundreds of thousands of clinicians use Heidi. Every clinician using Heidi gets hours back in their day. This is the starting point to build the product of your dreams: immense distribution, and a world of potential.
You’ll join a team that doesn’t celebrate fundraise milestones, imaginary valuations, and glossy PR. We live and breathe the pain of modern health systems, and won’t rest until we’ve exacted the change we’d like to see. We’re medicos, engineers, and designers who’ve felt (on every side of the equation) what non-care feels like - the moral and practical toll as a provider or receiver.
The bad news is it’s not an easy job - you will tear your hair out more than once. Modern technology company concepts don’t transpose unto healthcare easily. The lows are really low.
The good news is you will trust and admire everybody you work with, and rest easy knowing you’re doing THE defining work of your career. The highs are blindingly high.
Heidi will take care of you. We offer a generous $500 annual development budget and serious equity.
True A-players progress extremely fast here. The nature of the scale-up game is demanding, but we don’t track hours or micromanage - you’re trusted to perform. We value mental health and adapt to your schedule. We operate on outcomes > inputs, not process theatre. We all take the bins out, metaphorically and literally.
💡 Quick Summary
Seeking a career-building opportunity? The Product Manager, Models position is now open for candidates interested in the Back Office Jobs sector. This role in Sydney offers a professional environment and growth potential.
Requirement Snapshot: Candidates should possess basic communication skills, a proactive attitude, and the ability to work in a team. Experience in Back Office Jobs is a plus.