AI compensation guide: India, US and UK
Pay ranges by role family and seniority, with the source for every figure.
Skills move faster than titles
"AI engineer" can mean agents, fine-tuning or inference infrastructure.
The independent AI talent platform
Hire the people who build AI.
aigle maps the capabilities your AI work needs, evaluates specialists against that work, and delivers the engineers, researchers and leaders who fit. Permanent, contract or fractional.
Requirement to team
Illustrative scoping exampleAI-native scaleup · Series B
01Requirement
Launch an agentic assistant that resolves customer support requests end to end.
- Live within two quarters
- Production traffic from day one
- Hybrid team, European time zones
02Capabilities
C1LLM orchestration and tool usecoreC2Retrieval over product knowledge (RAG)coreC3Evaluation and guardrailscoreC4Latency and cost controlsupportingC5Tracing and monitoring (LLMOps)supportingC6Conversation and workflow designsupporting
03Evaluation
- AI Agent EngineerDesign review: a multi-step agent that must recover from tool failuresBuild exercise: a tool-calling agent with its own evaluation harness
- RAG EngineerDebugging retrieval quality on a messy, versioned document set
- LLMOps EngineerIncident walkthrough: tracing a cost spike and a quality regression
- AI Product ManagerProduct case: scoping the assistant and defining what success means
04Team and delivery
ModelContract-to-permanent team
Team5 people
Shortlist5–7 business days (target)
Select a role in the team to trace the capabilities it covers and how it is evaluated.Every engagement starts with a scoping call. Your map is built for your requirement.
Why aigle
We recruit AI talent. Nothing else.
Most firms that place AI specialists also place everyone else. Several of the best-known AI talent platforms now focus on supplying human data to model labs. aigle does one thing: helps organisations hire the people who design, build, ship and lead AI.
- Generalist recruitersAI is one desk among many.
- Self-serve marketplacesThe evaluation, and the risk, stay with you.
- Freelance platformsBuilt for transactional project work.
- Internal talent teamsCarry every function's hiring at once.
All functions
The AI talent problem
A CV shows keywords. The work shows depth.
Ordinary recruiting matches the words on a profile. AI hiring has to read the evidence behind them. Drag across the profile to see the difference.
What the work showsEvidence reviewed
Profile A · Senior AI Engineer
RAGShipped to production; repository and incident write-up
KubernetesOperated in production; on-call history
PythonProduction code across several repositories
PyTorchResearch code for one paper
AI agentsPrototype only; demo repository
EvaluationNo evidence found
Strong for retrieval and platform work. Not yet an agent engineer.
What changes the hire
Same keywords. A different engineer.
On keywords alone, Profile A matches an agent engineer role perfectly. The evidence says they would excel in retrieval and platform work, and would need time to grow into agents.
aigle evaluates the work behind every claim, so you meet the right person for the role you actually have.
How we evaluateThe evidence is scattered
Repositories, papers, model cards and incident histories rarely reach a CV.
Fluency is easy to show
Judgement about architecture and trade-offs takes a practitioner to test.
Research is not production
Most roles need a specific mix of the two, and the mix changes the hire.
72%
of employers worldwide report they cannot find the skilled talent they need.
No. 1
AI model and application development is now the hardest capability to hire for, worldwide.
Source: ManpowerGroup, Global Talent Shortage Survey 2026. Read the release
The aigle approach
A talent intelligence system, run by specialists
Six stages turn an AI hiring requirement into a team that fits it. Each stage leaves you with something you can read, question and keep.
- 01
Understand the objective
What the AI work must achieve, by when, and under which constraints.
You receive Scoping notes - 02
Define the capabilities
The objective translated into capabilities, seniority and team shape.
You receive Capability brief - 03
Map the talent
Where these people work today and how reachable they are.
You receive Talent map - 04
Evaluate the work
Role-specific assessment, reviewed by AI practitioners.
You receive Scorecards - 05
Present a shortlist
Human-reviewed candidates, each with the evidence behind them.
You receive Shortlist dossier - 06
Hire and measure
Interviews, offers, onboarding and a check-in after start.
You receive Placement review
On technology
Software helps us search wider and keep every record consistent. Specialists make the judgements, and your interview feedback recalibrates the search as it runs.
Services
One partner, seven ways to engage
Start from what you need. We recommend the model, and you can change it as the work changes.
Recommended model
Permanent AI hiring
Full-time hires for the roles your AI work is built around, run as a dedicated search when a role is urgent or confidential.
Good fit when
- The role is long-term and central to the roadmap
- You want one accountable partner for the search
- The bar is high and the market is thin
How it works
- Scoping call and capability brief
- Talent map and evaluated shortlist
- Offer support and a check-in after the start date
AI talent
The specialisms we recruit, and the skills behind them
Our recruiters work by AI sub-domain, so the person searching for your RAG engineer understands retrieval. Choose a family, then a role, to see the skills we assess.
43 roles across 6 families
Generative and agentic AI
The people who turn foundation models into products: agents that take actions, retrieval systems that ground answers, and the automation around them.
What we look for
- Shipped LLM features under real traffic, not only prototypes
- Evaluation habits: test sets, regressions, guardrails
- Sound trade-offs between models, cost and latency
Roles
Skills we assess
PythonLLM APIs: OpenAI, Anthropic, GeminiLangChain · LangGraphMCP and tool useVector databases: Pinecone, Weaviate, MilvusEvaluation and guardrailsAgent frameworks: CrewAI, AutoGenSelect a role to see the skills we assess for it
How we evaluate
One test cannot judge every AI professional
A research scientist, an agent engineer and a Head of AI prove themselves in different ways. Each role has its own evidence fingerprint, and we agree the scorecard with you before sourcing starts.
Evidence fingerprintIllustrative
Architecture and system designPrimary weight
Hands-on buildPrimary weight
Code and portfolio evidencePrimary weight
Research contributionLight weight
Production and operationsSupporting weight
Product and business judgementSupporting weight
Leadership and communicationLight weight
How we test it
A design review of a multi-step agent, a build exercise with an evaluation harness, and a walk through code they have shipped.
Illustrative weighting. Each search gets its own scorecard.
Who we work with
For organisations whose AI roadmap depends on a few hard-to-find people
- AI-native startupsFounding AI engineers and the first technical leaders, hired before the next round.First hires
- ScaleupsTeams for LLM products, agent systems and AI infrastructure as usage grows.Scale
- EnterprisesSpecialist hiring that works inside enterprise process and confidentiality.Process
- Global capability centresAI teams built across geographies, with India as a core talent market.Build
- Research organisationsScientists and research engineers for labs, robotics and autonomous systems.Research
Financial servicesHealthcare and life sciencesLLM companiesRetail and consumerAgentic AIManufacturing and industrialRobotics and autonomyAI SaaSAutonomous vehiclesEnterprise technologyEdTechPublic sector
Industries and sectors
Evidence and trust
What we commit to in writing
We publish client names, stories and results only with written approval. Until then, these are the standards every engagement is held to.
The replacement window on every permanent hire.
Scorecards you can read
Every shortlisted candidate comes with the evidence behind the recommendation: what was tested, how they did, and where the gaps are.
A 90-day replacement guarantee
If a permanent hire leaves within 90 days of starting, we search for a replacement as part of the same engagement.
Confidential by default
Searches can run without your name attached. Candidate information is shared only with their consent.
Market data with sources
Every salary and market figure we share states where it comes from and when.
Client stories appear here once clients approve them.Space reserved
Insights
Intelligence for AI hiring
Guides and data for people who hire AI specialists. The first editions are being prepared.
Hiring your first Head of AI
What the role owns in year one, how to scope it, and how to judge the shortlist.
How to evaluate an AI agent engineer
The design, build and evaluation skills that separate production agents from demos.
For AI professionals
Tell us your specialism once
Join the aigle network and hear only about roles that match what you do and how you want to work. Your details are shared with an employer only after you agree to a specific role.
- 01Share your specialism and a public profile.
- 02A specialist in your field reviews it.
- 03We contact you when a matching role opens.
Hire AI talent
Tell us what you are building
Share the requirement in a few lines. A specialist reviews it and replies with how we would approach the search.
- 01A specialist reads your requirement.
- 02We arrange a 30-minute scoping call.
- 03You receive a capability brief and a proposed approach.