San Francisco is where AI hiring happens first, so when a talent-search platform publishes six months of search data from that market, we read it as a leading indicator for Sydney, Melbourne and the wider APAC region. This brief covers searches run from January to June 2026. The figures are the platform's, not Sonitec research, and they describe what employers searched for, not what they hired. With that caveat, five things stand out.
1. Engineering demand hasn't dropped
Coding agents or not, Software Developer is still the most searched role by a wide margin. It accounts for 49% of top-ten role searches in the period. Full-Stack Engineer is next at 13% and Backend Engineer at 6%. If the prediction was that AI tooling would hollow out demand for people who write software, the search data from the city building most of that tooling does not show it yet.
2. AI experience is the fastest-growing ask
Employers were asked which backgrounds they wanted candidates to come from. AI and ML platform backgrounds went from 18% of requested backgrounds in the first half of 2025 to 24% in the first half of 2026. That is a 33% rise year on year, the biggest gain of any background in the data. The ask is not for AI researchers. It is for engineers who have shipped on top of AI and ML platforms, which is the same profile our Forward Deployed Engineer clients describe.
3. Applied AI Engineer breaks into the top 10
For the first time, Applied AI Engineer ranks among the ten most searched roles, at about 4% of top-role searches. The ranking is small next to Software Developer, and that is the point: AI-specific roles are gaining ground while core engineering stays on top. Teams are adding an applied AI seat beside the engineering team, not replacing the engineering team with it.
4. Skills matter more than sector
47% of searches named no industry background at all. Just over half still specified one, but almost half of employers were open to people from any sector as long as the skills were proven. That is a useful counterweight to the brief that insists on “must come from fintech” or “must have worked at a hyperscaler”. In a market this short of experienced AI engineers, the firms that screen on demonstrated work rather than a matching CV see more of the pool.
5. Inbound is losing value
AI-written CVs and cover letters make applications look the same. When every applicant reads well on paper, the job ad stops being a filter and the inbound pile stops being a shortlist. The response, visible in the search data, is that hiring is starting to look like executive search for roles well below executive level.
- Before: post a role, screen inbound applicants, judge on the CV.
- Now: identify the people you want, approach them directly, ask for proof of real work.
What this means if you are hiring in Australia
- Keep hiring engineers, and specify the AI exposure you need. The demand curve has not turned. What has changed is the extra line on the brief: which AI and ML platforms the person has shipped on, and what they built.
- Decide whether you need an Applied AI or Forward Deployed seat, not whether to rename an existing one. The San Francisco data shows a new role appearing beside core engineering. Our 2026 salary guide has the Australian bands for the AI seats.
- Drop the sector filter unless it is doing real work. Ask what the person built, on what stack, at what scale. That screens better than an employer name, and it widens a pool that is already thin in Australia.
- Plan for a direct search, not a job ad. If San Francisco employers are going direct for individual contributors, Australian employers competing for the same remote-capable people will need to as well. A direct approach with a specific brief gets replies that a posting no longer does.
- Ask for proof of work at the first conversation. A short technical walk-through of something the candidate shipped tells you more than any AI-polished document. Our voice interview exists for this reason.
How we use this data
We treat San Francisco search data as a signal, not a measurement of the Australian market. We cross-check it against the live roles we run and the Australian salary and job-ad data behind our salary guide before we change any advice. Tell us about the role and we come back with a search plan. A first shortlist takes at least 15 business days.
Source: talent-search platform data for San Francisco, January to June 2026, with January to June 2025 for the year-on-year comparison. Figures are as published by the platform and describe employer searches, not hires. First shared as the Sonitec Talent Market Brief, October 2026.