AI/Technical
Perpetuate’s AI/Technical practice is an Off-market search for VC-backed tech, AI-native and deep-tech companies: AI product engineers, forward deployed engineers, Staff and Principal engineers, AI product managers, data scientists, computer vision engineers and Heads of Engineering. Stephen Tung leads every search. A 20 to 30 minute founder brief, 2 to 3 personally assessed candidates in 5 to 10 days, and no job ads, ever.
Engineers and technical leaders placed since January 2025
%403x.png)
Which AI/Technical roles does Perpetuate place?
Perpetuate places eleven AI/Technical roles, from the engineer who ships the first LLM feature to the Head of Engineering who has to scale the team behind it. The people Stephen places are Operators: already embedded in serious work, rarely on the market, and only reachable through research-led, Craft-led approaches.
01
AI Product Engineers
From model capability to shipped LLM feature: prompts, evals, retrieval and orchestration. Placed at Atticus, Flowing Bee and Tendl.
02
Customer/Field Engineering and Senior Forward Deployed Engineering
The last mile inside a customer's data and workflow, turned into a working integration and a repeatable pattern for the core team.
03
Staff and Principal / Lead Engineers
Architecture, technical direction and the quality bar for a team they do not manage. Placed at Atticus, Zapier and Coviu.
04
AI Product (Technical) Management
What the model should do, for whom, and how you will know it worked. Trusted by engineers because they can reason about the system, not just the roadmap.
05
Data Scientists / Classical Machine Learning
Forecasting, optimisation, anomaly detection and experimentation, where a language model is the wrong tool for the job.
06
AI Data Scientists
Fine-tuning, evaluation design, synthetic data and multimodal pipelines. Placed the AI Lead at Sync Technologies.
07
VLM (Vision Language Model) / Computer Vision Engineers
Detection, segmentation, tracking and vision-language models on edge hardware, in robotics, construction and physical-AI environments.
08
VP / Head of Engineering
Delivery, hiring and engineering culture from one team to several. Placed at Lyro Robotics, Fleet Space, Deckard and Zapier.
09
Growth AEs
New revenue in a technical sale: discovery with a CTO, demos to engineers and multi-stakeholder deals for a company with no brand yet.
10
Head / VPs of Growth (Regional Sales)
A region's number and the team, motion and partnerships behind it, often as the first commercial leader in a founder-led sales culture.
11
AI (code-to-canvas) Design Engineers
The seam between design and front-end code: prototyping AI interactions in the codebase and shipping the interface, not a mock-up.
How an AI/Technical search runs
An AI/Technical search with Perpetuate starts with a 20 to 30 minute brief with the founder, CEO or CTO. Stephen then maps the market, approaches Operators directly and Off-market, and returns 2 to 3 personally assessed candidates within 5 to 10 days. Perpetuate manages both sides through to signed offer. No role is advertised at any point.
Day 1 · 20 to 30 minutes
The founder brief
One conversation with the decision maker: what you are building, the hard problem still open, the stage and the team. If the company is not yet somewhere an A-grade candidate would choose, Stephen says so on this call.
Days 1 to 3
Research and market mapping
Concentric-circle Signal recognition: who has built the closest thing to your problem, where, and who trained under them. Evidence of Alpha in what people have shipped, not keywords on titles.
Days 2 to 7
Off-market approach
Direct, personal notes to people who are not looking, each explaining why this company, this role and this moment fit their career. No job boards, no advertised roles, no mass outreach.
Days 5 to 10
A shortlist of 2 to 3
Every candidate has been interviewed by Stephen, assessed for technical depth, track record under constraint and Fit with your team, and knows why they are being introduced.
Through to offer
Both sides, one steward
Perpetuate runs diligence, references, compensation and timing for the company and the candidate together, and stays in the process until the offer is signed.
The best Operators paradoxically never need to apply.
They are embedded in serious work and entertain one or two outside conversations a year. Reaching them is research and Craft rather than volume: reading the Signal in what someone has actually built, recognising the Alpha in their experience, and being able to explain their own career back to them in a way they recognise. That is the whole method.
Stephen Tung, Founder, Perpetuate
The placement record behind the practice
Between January 2025 and September 2026 Perpetuate made 53 placements across 34 client companies, and 41% of those clients have hired through the firm more than once. Atticus alone has hired through Perpetuate five times.
53
placements across both practices
34
client companies
41%
of clients have hired again
“I wasn’t actively looking when Steve first approached me. His first note was different in a way I noticed immediately. He had taken the time to understand my work and academic background, and saw how my experience matched one particular company at one particular moment.”
SF
Stanislav Funiak
Carnegie Mellon PhD in Robotics and Machine Learning · former Facebook Research Scientist · employee #30 at Cerebras Systems · approached and placed Off-market by Stephen
Questions founders ask before an AI/Technical search
How do you hire a forward deployed engineer?
Hire for translation as much as for code. A forward deployed engineer sits inside a customer’s data and workflow and builds what makes your product real for that account, so the brief should name the customer environments they will face. Perpetuate finds them Off-market among engineers and scientists already doing high-context integration work, and shortlists 2 to 3 within 5 to 10 days.
How do AI-native startups reach Staff engineers who aren’t looking?
Job ads will not reach them. Staff and Principal engineers who are shipping serious systems do not apply anywhere; they respond to a specific, well-researched note about a problem they would want to own. Perpetuate maps who has built the closest thing to your system, approaches them directly, and explains why this company and this moment fit. Atticus’s Staff Software Engineer (LLM) was found this way.
What is the difference between an AI product engineer and a machine learning engineer?
An AI product engineer builds product on top of models: prompts, retrieval, evals, orchestration and the user-facing feature. A machine learning engineer or data scientist builds and trains the models themselves, from classical forecasting to fine-tuning. Perpetuate places both, and treats them as different searches with different Signal, which is why the practice lists them as separate roles.
How much of a founder’s time does an AI/Technical search take?
About 20 to 30 minutes for the brief, then a short conversation when the shortlist arrives 5 to 10 days later, then interviews with 2 to 3 candidates you actually want to meet. Perpetuate handles research, approach, assessment, scheduling and both sides of the offer, so the founder’s time goes into deciding rather than sourcing.
Does Perpetuate place outside Australia?
Yes. Perpetuate is Zapier’s exclusive ANZ hiring partner, has placed a Senior Frontend Engineer in New Zealand, and runs searches for companies expanding into the UK and US. Australian and New Zealand technical talent is the core market; the candidates and the companies are increasingly global.
Search for New Zealand companies →The engineer you need is shipping somewhere else right now.
Twenty to thirty minutes with Stephen is the whole brief. Two to three personally assessed, Off-market candidates follow within 5 to 10 days, and Perpetuate runs both sides through to signed offer.
Email Stephen to brief a searchMessage Stephen on LinkedIn →