Technical AI talent forRLHF & Alignment

On-Demand Experts To Train Your Frontier Models.

Technical AI expertise.

On-demand human expertise powering frontier AI through RLHF, model evaluation, programming expertise, and creative intelligence. Published hourly rates, vetted specialists, shortlists in 48 hours.

0 expert roles with published rates • 0-hour shortlists • 0+ countries

0
Expert Roles, Rates Published
0h
Shortlist Turnaround
0
Technical Domains
0+
Countries Sourced

Specialists from teams like these are in our network

What We Provide

The technical workforce behind frontier AI.

RLHF & Human Feedback
Reward modelling, preference data, and written critique from specialists who understand what a training signal has to look like to be useful.
Reward ModellingPreference DataCritique WritingReviewer Calibration
Model Evaluation
Structured grading of reasoning, factuality, and instruction-following against your rubric, with the justification your training loop can learn from.
Rubric GradingReasoning TracesFactuality ReviewBenchmark Design
AI Safety & Red Teaming
Adversarial prompt testing, jailbreak discovery, hallucination detection, and robustness testing measured against your published safety policy.
Adversarial PromptingJailbreak TestingHallucination DetectionRobustness Testing
Code Evaluation & Programming Experts
Senior engineers who write reference solutions and grade model-generated code across Python, TypeScript, JavaScript, Rust, Go, Java, and C++.
Reference SolutionsDiff GradingRepo-Scale TasksSecurity Review
AI Data & Datasets
Dataset design, curation, contamination checks, and quality control with agreement scoring and an audit trail on every batch.
Dataset DesignData CurationQC AuditsContamination Checks
Creative Intelligence
Directors, cinematographers, editors, and sound designers for generative creative workflows and purpose-shot video and audio data.
Generative WorkflowsCreative EvaluationVideo CaptureAudio & Sound

Who We Work With

Frontier AI LabsFoundation Model TeamsPost-Training & RLHF TeamsAI Safety & Evaluation OrgsAgent & Tooling CompaniesVideo & Audio Model TeamsAI Product Engineering Teams
Frontier AI LabsFoundation Model TeamsPost-Training & RLHF TeamsAI Safety & Evaluation OrgsAgent & Tooling CompaniesVideo & Audio Model TeamsAI Product Engineering Teams
Frontier AI LabsFoundation Model TeamsPost-Training & RLHF TeamsAI Safety & Evaluation OrgsAgent & Tooling CompaniesVideo & Audio Model TeamsAI Product Engineering Teams
Frontier AI LabsFoundation Model TeamsPost-Training & RLHF TeamsAI Safety & Evaluation OrgsAgent & Tooling CompaniesVideo & Audio Model TeamsAI Product Engineering Teams

Roles & Rates

Expert roles, with the rate published on each.

Every role carries a published hourly rate. You are billed for hours worked, with no sourcing fee and no mark-up on top.

RLHFRLHF & Alignment

RLHF Engineer

Builds and runs reinforcement learning from human feedback pipelines: reward modelling, preference data collection, grading rubrics, and post-training evaluation loops.

Reward ModellingPreference DataDPO / PPOPost-Training Evals
Hourly rate$80–$120/hr

Remote · Contract · Full-time available

EVALModel Evaluation & QA

AI Model Evaluator

Grades model output against your rubric across reasoning, factuality, and instruction-following, with the written justification your training loop needs.

Rubric GradingReasoning TracesFactuality ReviewEval Reporting
Hourly rate$45–$75/hr

Remote · Contract · Flexible hours

SAFEAI Safety & Red Teaming

AI Safety Specialist

Defines safety policy, builds refusal and over-refusal taxonomies, and measures model behaviour against your published safety commitments.

Safety PolicyRefusal TaxonomyHarm CategorisationPolicy Evals
Hourly rate$75–$115/hr

Remote · Contract · Flexible hours

PLEEngineering & Data

Programming Language Expert

Senior engineers in a single language who write reference solutions and grade model code on idiom and correctness — Python, TypeScript, JavaScript, Rust, Go, Java, and C++.

PythonTypeScriptRustGoJavaC++JavaScript
Hourly rate$50–$90/hr

Remote · Contract · Flexible hours

REDAI Safety & Red Teaming

Red Team Evaluator

Attacks the model on purpose: adversarial prompting, jailbreak discovery, hallucination probing, and robustness testing under distribution shift.

Adversarial PromptingJailbreak TestingHallucination ProbingRobustness Testing
Hourly rate$60–$100/hr

Remote · Contract · Project-based

ADSEngineering & Data

AI Data Scientist

Owns dataset design, sampling strategy, contamination checks, and the analysis that tells you whether a training run actually moved the metric.

Dataset DesignSampling StrategyContamination ChecksEval Analysis
Hourly rate$70–$110/hr

Remote · Contract · Full-time available

ACSModel Evaluation & QA

AI Coding Specialist

Builds coding tasks, unit-tested benchmarks, and repository-scale problems that measure whether a model can actually ship working software.

Benchmark DesignRepo-Scale TasksUnit Test HarnessesSWE-Style Evals
Hourly rate$55–$95/hr

Remote · Contract · Full-time available

PEXRLHF & Alignment

Prompt Engineering Expert

Designs prompt suites, system instructions, and task taxonomies that surface real model behaviour instead of best-case demos.

Prompt SuitesSystem InstructionsTask TaxonomyFailure Discovery
Hourly rate$50–$85/hr

Remote · Contract · Flexible hours

CAICreative AI & Media

Creative AI Specialist

Directs generative image, video, and audio workflows end to end, and evaluates creative model output against a brief a client would actually sign off.

Generative WorkflowsCreative EvaluationStyle DirectionModel Comparison
Hourly rate$45–$80/hr

Remote · Contract · Project-based

In The Network

Technical AI specialists, available by the hour.

RLHF EngineerAI Model EvaluatorAI Safety SpecialistRed Team EvaluatorPrompt Engineering ExpertHuman Feedback ExpertAI QA TesterCode Reviewer
RLHF EngineerAI Model EvaluatorAI Safety SpecialistRed Team EvaluatorPrompt Engineering ExpertHuman Feedback ExpertAI QA TesterCode Reviewer
RLHF EngineerAI Model EvaluatorAI Safety SpecialistRed Team EvaluatorPrompt Engineering ExpertHuman Feedback ExpertAI QA TesterCode Reviewer
RLHF EngineerAI Model EvaluatorAI Safety SpecialistRed Team EvaluatorPrompt Engineering ExpertHuman Feedback ExpertAI QA TesterCode Reviewer
AI Data ScientistProgramming Language ExpertAI Coding SpecialistCUDA & Kernel EngineerAgent & Tooling EngineerTraining Data SpecialistCreative AI SpecialistCinematographer / DOP
AI Data ScientistProgramming Language ExpertAI Coding SpecialistCUDA & Kernel EngineerAgent & Tooling EngineerTraining Data SpecialistCreative AI SpecialistCinematographer / DOP
AI Data ScientistProgramming Language ExpertAI Coding SpecialistCUDA & Kernel EngineerAgent & Tooling EngineerTraining Data SpecialistCreative AI SpecialistCinematographer / DOP
AI Data ScientistProgramming Language ExpertAI Coding SpecialistCUDA & Kernel EngineerAgent & Tooling EngineerTraining Data SpecialistCreative AI SpecialistCinematographer / DOP

How It Works

From brief to working specialists in four steps.

01
Tell us the expertise
The role, the technical depth you need, the hours, and the start date. One brief, no job spec required.
02
Shortlist in 48 hours
We send interviewed specialists with verified technical history and a confirmed hourly rate for each.
03
Trial the work
Start with a paid trial batch measured against your rubric. If the output misses the bar, you do not continue.
04
Scale or stop
Add specialists as the training run grows, or end the contract when the project does. Billing stays hourly.

Why Fradle

An expert network built for frontier AI teams.

01
Built for frontier AI, not general hiring
Every role on this platform exists because a model team needed it: RLHF, evaluation, red teaming, code review, and creative intelligence.
02
Verified technical depth
Specialists are checked against shipped work, published repositories, and a graded task in their own domain before they reach your shortlist.
03
Published hourly rates
Every role carries its rate on this site. You are billed for hours worked, with no sourcing fee and no mark-up on top.
04
48-hour shortlist
Most briefs return interviewed specialists within two business days, including for narrow evaluation and safety work.
05
Replace at no charge
If a specialist does not meet your quality bar during the trial batch, we replace them at no additional cost.
06
Contracts and payroll handled
We hold the contracts, handle payment, and manage compliance across 30+ countries so your team does not have to.

What People Say

From model teams and the specialists themselves.

"We needed RLHF engineers who had actually shipped a reward model, not reviewers handed a rubric. Fradle had three working inside our pipeline in under a week."

ER
Dr. Elena Rostova
Post-Training Lead, Frontier Model Team

"Their red team evaluators found jailbreak classes our internal suite had never covered. The reports came back structured enough to action the same day."

SJ
Sarah Jenkins
Head of AI Safety Evaluation

"Code reviewers who could tell a plausible diff from a correct one changed our coding benchmark scores overnight. Billing by the hour meant we scaled it exactly as far as we needed."

MV
Marcus Vance
Head of Model Evaluation

"For our video model we needed a real DOP and a real sound designer running purpose-shot capture. The footage quality showed up directly in the training set."

DC
David Chen
Data Lead, Video Model Startup

"I grade reasoning traces in my own specialty a few hours a week at a rate that makes it worth my time. The scope was clear from day one."

AI
Amina Idris
AI Model Evaluator, Fradle Network

Frequently Asked Questions

What we do, and what we do not.

Technical experts for frontier AI teams: RLHF engineers, AI model evaluators, human feedback experts, AI safety specialists, red team evaluators, AI QA testers, code reviewers, AI coding specialists, programming language experts, AI data scientists, and training data specialists. Creative AI and media professionals are a secondary offering for teams doing generative creative work and video or audio data capture.

Tell Us What You Are Training

Get the technical expertise your model needs
in 48 hours.

Send us the role, the technical depth, and the hours. You get interviewed specialists with confirmed rates. Or join the network if you are the specialist.