Job market analytics for engineering roles resistant to AI replacement. Roles emphasising physical/hands-on work, precision trades, field work, robotics, and hardware.
Roles tagged as AI-proof score ≥2 keyword hits across: hardware, field work, robotics, precision trades, safety/compliance, and physical manufacturing. These domains require on-site presence, physical dexterity, or domain-specific judgment resistant to AI automation.
| # | Company | Jobs | Locations |
|---|---|---|---|
| 1 | Jobster | 67 | Nederland, Landgraaf, Wageningen |
| 2 | Haskoning | 23 | Eindhoven, Amersfoort, Delft |
| 3 | Trinamics | 19 | Eindhoven, Dordrecht, Breda |
| 4 | Proxify | 13 | EMEA |
| 5 | Playrix | 12 | Netherlands, European Union, EMEA |
| 6 | Jobgether | 11 | Netherlands, EMEA |
| 7 | OverheidZZP | 10 | Houten, Rotterdam, ’s-Hertogenbosch |
| 8 | Twine | 9 | European Economic Area |
| 9 | Doghouse Recruitment | 8 | Amsterdam, Hilversum, Utrecht |
| 10 | Athora Netherlands | 8 | Amsterdam |
| 11 | Booston.io | 8 | Oss, Zwolle, Utrecht |
| 12 | Looper Engineering | 8 | Eindhoven, Breda, Rotterdam |
| 13 | Capgemini | 7 | Utrecht |
| 14 | Cyso Cloud | 7 | Alkmaar |
| 15 | Lumenalta | 7 | Eindhoven, Amsterdam, European Economic Area |
| 16 | Xccelerated | 6 | The Randstad, Eindhoven Area, Utrecht |
| 17 | Spilberg | 5 | North Holland, Arnhem, Utrecht |
| 18 | Sweco | 5 | Amsterdam, Apeldoorn, Zwolle |
| 19 | Databricks | 5 | Amsterdam |
| 20 | WR.nl Solliciteren | 5 | Groningen, Zeewolde, Uden |
You don't have to leave software. You have to go deeper — into domains where AI writes code, but humans carry the accountability, the infrastructure, and the liability. These paths combine your existing skills with layers of AI-resistant context.
Build the internal developer platforms, internal tooling, and infrastructure abstractions that everyone else uses. When you own the platform, you own the leverage. AI can't replace the person who built the environment it runs in.
AI writes YAML — you build the system that runs it
SREs own the reliability contract between engineering and the business. On-call, incident response, SLA enforcement, and error budgets require judgment calls with real financial consequences. AI assists but can't take the 3am call.
AI-assisted alerting — human owns the outcome
Shift security left into the pipeline, own the threat model, do penetration testing, and design zero-trust architectures. Security engineers sign off on architecture — that's legal accountability AI can't carry.
Humans sign security reviews — not AI models
Own the tools that other developers use: internal CLIs, scaffolding generators, IDE plugins, code quality dashboards, and CI/CD templates. You determine how everyone else ships code — high leverage, high ownership.
Build the AI tools — don't be used by them
Build the data pipelines, feature stores, model training infrastructure, and deployment systems that power AI — rather than being replaced by it. Data engineering owns the data contract — essential regardless of what models run on top.
Pipelines outlive models — own the plumbing
Work at the layer below application code: distributed systems, consensus protocols, storage engines, networking kernels. These require formal reasoning about correctness, performance, and failure modes that AI can't reliably produce.
Correctness proofs > autocomplete