Mercor vs Harnham: full comparison for 2026
Quick verdict
Mercor (3.7/5) edges ahead of Harnham (3.7/5) overall. Mercor is the better choice for AI labs and research teams that need specialist contractors in large numbers. Harnham is the stronger option for companies hiring permanent data or ML staff in the UK or U.S. through a specialist agency. The right choice depends on your project size, budget, and required tech stack.
Mercor vs Harnham: head-to-head summary
| Criterion | Mercor | Harnham |
|---|---|---|
| Founded | 2023 | 2006 |
| HQ | San Francisco, California, USA | London, United Kingdom |
| Team size | 300–400 staff; large contractor network | 100–500 |
| Rating | 3.7 / 5 | 3.7 / 5 |
| Primary differentiator | AI-run interviews and matching built for high-volume expert hiring | Twenty years of recruiting only in data and analytics |
| Pricing model | Contractor rate plus platform fee (about 30%, Sacra estimate) | Placement fee for permanent hires; contractor day or hourly rates; rates on request |
| Min. engagement | Not published | Not published |
| Primary tech stack | Python, PyTorch, OpenAI | Python, SQL, Spark |
| Industries served | AI labs, Technology, Finance, Legal, Healthcare | Financial services, Retail, Healthcare, Media, Technology |
Mercor vs Harnham: overview
Mercor
Mercor was founded in San Francisco in 2023 and uses AI interviews to screen applicants. It raised money at a $10 billion valuation in October 2025, mainly on the strength of supplying experts to AI labs for model training and evaluation. Product teams can hire engineers through it, but the platform is built for volume, and Sacra estimates its fee at about 30% of contractor pay. If you want two senior ML engineers for a year-long roadmap, look elsewhere. A firm that employs and manages its people fits that job better.
Harnham
Harnham has recruited for data and analytics roles since 2006 from London, with offices in the U.S. including New York and San Francisco. It places data engineers, data scientists and ML engineers on contract or permanent terms and runs a graduate training arm, Rockborne. As a recruitment agency, it screens through consultants who specialise in data hiring rather than through practising engineers, and contractors are not managed after placement the way a staffing firm's employees are. That makes it better for permanent hires than for managed augmentation.
Services and capabilities: Mercor vs Harnham
| Capability | Mercor | Harnham |
|---|---|---|
| ML engineers | ✓ | ✓ |
| LLM / GenAI engineers | ✓ | ✗ |
| AI agent developers | ✗ | ✗ |
| MLOps engineers | ✗ | ✗ |
| Computer vision engineers | ✗ | ✗ |
| NLP engineers | ✗ | ✗ |
| Data engineers | ✗ | ✓ |
| Engineer-led technical screen | ✗ | ✗ |
| Fractional / part-time experts | ✗ | ✗ |
| Trial before commitment | ✗ | ✗ |
| Nearshore time-zone overlap | ✗ | ✗ |
| Direct hire option | ✗ | ✓ |
Tech stack comparison: Mercor vs Harnham
| Framework / platform | Mercor | Harnham |
|---|---|---|
| PyTorch | ✓ | N/A |
| TensorFlow | N/A | N/A |
| LangChain | N/A | N/A |
| Hugging Face | N/A | N/A |
| OpenAI | ✓ | N/A |
| AWS | ✓ | ✓ |
| Azure | N/A | ✓ |
| Google Cloud | N/A | N/A |
| Databricks | N/A | ✓ |
| Kubernetes | N/A | N/A |
Pricing comparison: Mercor vs Harnham
| Criterion | Mercor | Harnham |
|---|---|---|
| Minimum engagement | Not published | Not published |
| Engagement models | Freelance contract | Direct hire, Contract-to-hire, Freelance contract |
| Rate transparency | Not public | Not public |
| Price tier | Mid-market | Mid-market |
Target audience comparison: Mercor vs Harnham
| Dimension | Mercor | Harnham |
|---|---|---|
| Best company size | Startup to mid-market | Startup to mid-market |
| Best industries | AI labs, Technology, Finance | Financial services, Retail, Healthcare |
| Best use cases | Hiring dozens of domain experts to evaluate a model, Adding a contract ML engineer for a research sprint | Hiring a permanent head of data science in London, Placing a contract data engineer for six months |
| Typical project type | Freelance contract | Direct hire |
Mercor vs Harnham: pros and cons
| Mercor | |
|---|---|
| + | Fast access to a large pool of specialists |
| + | Well funded |
| + | Experienced with AI-lab evaluation and training work |
| - | AI interviews, not engineers, do the first screen |
| - | Fee of about 30% adds up over a long engagement |
| - | Founded in 2023, so a short track record with product teams |
| Harnham | |
|---|---|
| + | Long specialist history in data recruiting |
| + | Offices in the UK and several U.S. cities |
| + | Both contract and permanent hiring |
| - | Screening by recruitment consultants, not engineers |
| - | Contractors are not managed after placement |
| - | Headcount estimates vary |
Who should choose Mercor?
A typical fit: hiring dozens of domain experts to evaluate a model.
AI-run interviews and matching built for high-volume expert hiring. Minimum engagement is not publicly disclosed. Works best with clients in AI labs, Technology, Finance, Legal, Healthcare.
Who should choose Harnham?
A typical fit: hiring a permanent head of data science in London.
Twenty years of recruiting only in data and analytics. Minimum engagement is not publicly disclosed. Works best with clients in Financial services, Retail, Healthcare, Media, Technology.
Decision matrix: Mercor vs Harnham
| Your situation | Recommended choice |
|---|---|
| You want a working engineer, not a recruiter, to run the technical screen | Neither documents an engineer-led screen; run your own technical interview |
| You need one specialist for a few days a week | Neither advertises part-time experts; ask about reduced hours |
| You need several engineers working as one team | Neither lists dedicated teams; check team size before signing |
| You want to test an engineer before committing | Neither publishes a trial; negotiate a short first term |
| Your budget is at the lower end | Compare: Mercor (Not published) vs Harnham (Not published) |
| Your team works U.S. hours | Neither lists Latin American engineers; confirm overlap hours in the contract |
| You may want to hire the engineer permanently later | Harnham |
Use case fit: Mercor vs Harnham
| Use case | Mercor fit | Harnham fit | Winner |
|---|---|---|---|
| Hiring dozens of domain experts to evaluate a model | Strong | Strong | Both equally |
| Adding a contract ML engineer for a research sprint | Strong | Limited | Mercor |
| Hiring a permanent head of data science in London | Strong | Strong | Both equally |
| Placing a contract data engineer for six months | Limited | Strong | Harnham |
Verdict: Mercor vs Harnham
Mercor (3.7/5) is the stronger overall choice for most AI Engineer Staffing projects. AI-run interviews and matching built for high-volume expert hiring.
Harnham (3.7/5) is worth a look if you need placing a contract data engineer for six months. If your situation matches that, Harnham is a competitive option.
Related comparisons
Mercor vs Harnham FAQ
Is Mercor better than Harnham?
Mercor (3.7/5) scores higher overall, but "better" depends on your use case. Mercor's strongest advantage: fast access to a large pool of specialists. Harnham's strongest advantage: long specialist history in data recruiting.
How do Mercor and Harnham differ in pricing?
Mercor uses contractor rate plus platform fee (about 30%, sacra estimate) pricing. Harnham uses placement fee for permanent hires; contractor day or hourly rates; rates on request pricing. Neither firm publishes a full rate card; a discovery call is required for project-specific quotes.
Which is better for enterprise: Mercor or Harnham?
Mercor is the larger team and typically the better enterprise-scale choice. For very large programmes, verify team size and compliance coverage directly with each company before shortlisting.
What are the main differences between Mercor and Harnham?
Mercor's primary differentiator is: AI-run interviews and matching built for high-volume expert hiring. Harnham's primary differentiator is: twenty years of recruiting only in data and analytics. They also differ in team size (300–400 staff; large contractor network vs 100–500), minimum engagement (Not published vs Not published), and primary industries served (AI labs, Technology vs Financial services, Retail).
Verify all details directly with each company before making a decision.