How to verify an AI engineer’s credentials
Check a credential where it is issued, not where it is advertised. A badge a third party awards, a track record you can audit, and production systems you can ask about tell you far more than a percentile a platform grades for itself.
What does “vetted” actually mean?
It depends on who did the checking. Across AI-staffing marketplaces it means one of three things:
| Kind of claim | Example | Can you check it? |
|---|---|---|
| Self-graded percentile | A platform’s own “top 3%” of applicants | No: the platform grades its own roster |
| Speed as proof | A promise of fast matching | No: speed says nothing about skill |
| Third-party badge | Upwork Expert-Vetted, shown on a public profile | Yes: issued and maintained by someone else |
Self-graded claims are not necessarily false. They are just each provider’s own statement, so the first real test of the engineer becomes your project.
Which four checks should you run?
All four take under ten minutes from public profiles and a short call, with no technical interview:
- A verified vetting tier. A badge the platform awards after its own review, not a line on a CV.
- A track record you can audit. A Job Success Score and contract history across years, with reviews written by clients.
- Production, not notebooks. Ask what runs inside a real company because of them. Notebooks show someone can model; production references show they can ship.
- Business-terms communication. Can they explain the trade-off they made, and what it cost or saved, without hiding behind jargon?
How do you check an Upwork Expert-Vetted badge?
Open the engineer’s public Upwork profile. The Expert-Vetted badge, the Job Success Score and the contract history are displayed there, awarded and maintained by Upwork rather than by the freelancer or an agency reselling them. Expert-Vetted marks the top 1% of talent on the platform, screened by Upwork’s own talent team. If a provider cannot point you to a profile like that, its vetting is a claim you cannot check.
What should you ask about production work?
- What is running today because of you, and for how many users or employees?
- What broke after launch, and how did you find out?
- How is it evaluated and monitored, and who gets paged?
- What did you hand over when you left, and where does it live?
What are the red flags?
- A percentile with nothing behind it you can open and read.
- A portfolio of notebooks and demos with no system in production.
- “Experienced with LLMs” with no named deployment or scale.
- Speed offered as the main proof of quality.
What does a verifiable AI engineer look like?
Hold every candidate to the same standard, including us. Here is how Cat Yung, Super Cat’s founder, scores on the four checks:
- Vetting tier: Expert-Vetted on Upwork, top 1% platform-wide (source: Upwork profile, October 2026).
- Track record: Top Rated Plus, 100% Job Success Score across 21 AI agent, RAG and LLM engagements and 1,152+ hours (source: Upwork profile, October 2026).
- Production: a 24/7 AI voice agent at a 5,000-employee US enterprise and enterprise RAG inside a Big 4 accounting firm, written up on the portfolio.
- Professional history: on LinkedIn.
What can a credential not tell you?
A badge proves who checked an engineer, not that they fit your problem. Even a verified engineer needs a scoped first piece of work, access to your data and an owner on your side. Treat the credential as the reason to start, and the first few weeks as the real interview.
Frequently asked questions
Is an Upwork Expert-Vetted badge worth checking?
Yes. It marks the top 1% of talent on Upwork, screened by Upwork’s own talent team rather than self-reported, and it is checkable on the public profile in seconds. Combine it with the Job Success Score and named production work.
Is “top 3%” a meaningful credential?
It is the platform’s own claim about its own applicants, so you cannot check it independently. Look for signals a third party issues and maintains instead.
How long does it take to verify an AI engineer?
The four checks take under ten minutes from public profiles and a short call: vetting tier, auditable track record, production evidence, and business-terms communication.
Do I still need a technical interview?
Not to verify credentials. You do still need a scoped first piece of work, because a credential proves who checked the engineer, not that they fit your problem.
Engineers whose vetting a third party audited.
Platforms, rates and a shortlist.
Marketplaces and specialists, compared.
The credentials above, in full.
About the author
Cat Yung is the founder of Super Cat Technology, an AI agent engineering team in Hong Kong and London, and works as an AI expert, data scientist and fractional CTO. Over a decade in production AI, NLP and machine learning; Expert-Vetted top 1% on Upwork with a 100% Job Success Score (source: Upwork profile, October 2026).