AI engineers
AI engineering splits into four distinct jobs that hiring managers are routinely asked to fill with one person. This page covers the combined role, and points to the four specialisations underneath it.
- Specialisations
- GenAI, agentic AI, ML, AI infrastructure
- Common stack
- Python, model APIs, vector and hybrid search, orchestration
- Cloud AI services
- AWS, Azure and Google Cloud AI platforms
- Engagement shapes
- Single engineer, embedded pod, fixed-scope assessment
- Typical duration
- Six weeks to nine months
- Evaluation
- Every engagement includes an evaluation harness
CloudOps Network matches AI engineers across four specialisations: generative AI, agentic AI, machine learning, and AI infrastructure. The combined AI engineer role suits teams that need one person spanning retrieval, orchestration and evaluation. Larger programmes usually need two or three of the specialisations working together.
What AI engineers do on our engagements
Generative AI
Retrieval quality, context engineering, output evaluation and cost per request.
GenAI engineersAgentic AI
Tool-calling systems, state, retries, human handoff and traceability.
Agentic AI engineersMachine learning
Training, serving, feature pipelines, drift and monitoring.
ML engineersAI infrastructure
GPU capacity, inference serving, gateways, throughput and unit cost.
AI infrastructure engineersData foundations
The pipelines and quality work that most AI projects discover they needed first.
Data engineersFull implementation
The combined engagement, from assessment through to production handover.
AI implementationThe question that filters AI engineers fastest
Ask a candidate how they know whether a change made the system better. Engineers who have shipped AI to production answer with an evaluation method. Engineers who have only built demos answer with an anecdote.
Everything else follows from that. Without evaluation, an AI system cannot be maintained, only redecorated.
Evaluation before architecture
If they reach for the harness first, they have been burned before. That is what you want.
Data quality realism
Good engineers ask about your documents early, because that is usually where the quality ceiling sits.
Cost awareness
Cost per request should be a design constraint they raise, not a question you have to ask.
Failure design
What the system does when it does not know is a product decision, and they should have opinions about it.
A sample of the bench
Availability changes weekly. This is the shape of who is in the network, not a live list.
Illustrative profiles drawn from the current roster. Names and clients are withheld until a mutual NDA is in place; the shortlist you receive names real people with real references.
Questions about hiring AI engineers
Do I need a GenAI engineer or an ML engineer?
If you are building on top of existing models, you want a GenAI or agentic engineer. If you are training, tuning or serving your own models against your own data, you want an ML engineer. Teams often brief the second and need the first.
Can one AI engineer handle a whole project?
For a bounded system, yes. Larger programmes usually need a GenAI engineer plus a data engineer, and an AI infrastructure engineer once self-hosted inference is involved.
Do your engineers work with self-hosted or open-weight models?
Yes, including the GPU capacity planning, serving architecture and throughput work that comes with running inference yourself.
How do you verify AI experience?
The technical screen walks through a production system the engineer built: evaluation approach, failure modes, latency and cost decisions. Demo-only experience does not survive that conversation.
Can they work under our data residency constraints?
Yes. Residency, processing boundaries and model provider restrictions are part of intake, and a data processing agreement is standard contracting.
Tell us what you are building
Describe the work in a few lines. We come back with a shortlist of engineers who have done it before, their availability, and a rate range. No account, no platform fee, no obligation to hire.
Most briefs get a shortlist within three business days.