How we vet experts
Before anyone works on your project, they complete an in-depth AI interview with AI Screenr, our AI recruiter, and a senior Jobrela specialist reviews the evidence. You get experts who have proven what they know, not just listed it on a CV.
Every expert passes an AI interview and a human review
Four steps between a promising CV and an expert on your project.
Profile and requirements
We define what the project needs, from must-have skills and their level to knockout criteria and the language level.
AI interview
The candidate does a structured voice interview of 15 to 60 minutes with the AI interviewer, at any time that suits them, in their own language.
Scored report
Every answer is scored against a rubric, with evidence quotes from the transcript.
Human decision
A senior Jobrela specialist reviews the report and the transcript and makes the final decision. AI recommends, people decide.
What a Java developer interview report looks like
This is the report our clients receive after every AI interview. The candidate is invented, so no real person's data is shown, but the structure is exactly the same.
Summary
The candidate showed solid, hands-on Spring Boot and microservices experience and explained real design decisions with concrete examples. All knockout criteria are met. The recommendation is “Yes” rather than “Strong yes” because Kubernetes production experience, a must-have for this role, was not shown, and answers on concurrency stayed at the level of definitions.
An experienced backend developer who has built and run Spring Boot services in production, with a good grasp of API design, transactions and resilience patterns. The main gaps are hands-on Kubernetes operations and depth in Java concurrency. A strong fit for the role if the team can support the Kubernetes side during onboarding.
Knockout criteria and must-haves
Right to work in the EU
The candidate is an EU citizen and needs no work permit.
"I’m a Czech citizen, so no permit needed."
Availability
Can start within the required four weeks.
"My notice period is one month, I could start at the beginning of December."
At least 4 years of Java backend experience
Has about six years of professional Java backend development.
"About six years now, mostly Spring Boot services for e-commerce and payments."
Spring Boot (advanced)
80 %Built and maintained several Spring Boot services in production and explained configuration and testing choices in detail.
"We moved all services to Spring Boot 3 last year, the biggest change was the Jakarta namespace and native observability."
REST API design and microservices (intermediate)
75 %Described versioning, idempotency and service boundaries from own experience.
"Every payment request carries an idempotency key, so a retry never charges twice."
SQL and JPA/Hibernate (intermediate)
70 %Knows how to find and fix N+1 queries and when to drop down to native SQL.
"The N+1 problem showed up in the order history, we fixed it with an entity graph."
Kubernetes in production (intermediate)
60 %Only deployed with a prepared Helm chart; no experience with troubleshooting or scaling workloads.
"The platform team owned the cluster, we just changed the values file."
Scores in ten areas
Explained technical decisions clearly and in a logical order, with occasional long detours.
"We split the monolith along the payment and order boundaries first, because those two teams kept blocking each other."
Answers stayed on the question and were tied to the candidate’s own projects.
"In our case the retry happened in the gateway, so the service itself had to be idempotent."
Strong practical knowledge of Spring Boot, JPA and REST API design; good understanding of transactions and resilience patterns.
"We used a circuit breaker around the card provider, and a dead-letter queue for anything that still failed after three retries."
Walked through a production incident step by step, from symptoms to root cause and fix.
"The connection pool was exhausted, so first I looked at which queries held connections the longest."
Has built exactly the kind of services the role needs and has worked in similar team setups.
Calm and confident, and openly said where their knowledge ends.
"Honestly, I have not tuned that myself, the platform team did."
Gave one good example of mentoring a junior developer, but little about handling disagreement in the team.
Most answers were complete; some technical answers stopped at the definition without trade-offs.
Fluent professional English, comfortable discussing technical topics in depth.
Strong on Spring Boot and API resilience, thinner on concurrency and transaction propagation.
Prepared technical questions
B1. What are the main differences between Spring Boot and the traditional Spring Framework?
+ Explained auto-configuration with a concrete example from a real migration.
- Did not mention how to exclude an auto-configuration that gets in the way.
B2. What does the volatile keyword do, and when would you use it?
+ Gave a correct, practical example with a shutdown flag.
- Could not explain why count++ on a volatile field is still not thread-safe.
B3. How would you design a payment API that stays correct when clients retry and downstream services fail?
+ Answered from a real production system, with numbers and trade-offs.
- Did not cover how events stay consistent with the database.
B4. What are common pitfalls with @Transactional in Spring?
+ Knew the self-invocation problem from experience.
- Unsure about propagation levels.
Hands-on Kubernetes operations and deeper Java concurrency.
- Hands-on Kubernetes troubleshooting and scaling.
- Trade-offs between concurrency approaches and the Java memory model.
- Transaction propagation levels and when to use them.
- An example of resolving a technical disagreement in the team.
Suggested next step
Invite to a technical interview with the team lead. Focus on a short Kubernetes troubleshooting scenario and a concurrency exercise, for example a thread-safe cache.
How ten Java developer candidates compare
Every candidate for a role answers the same questions, so you can rank them, compare them area by area and see the strengths and risks of the best three. Invented candidates, real structure.
Ranking
You decide the criteria. These ten are the default, but you can add, remove or reweight any of them. Each cell is coloured by its score; a star marks a value well above the group, a triangle one well below it. Coverage shows how much of the prepared questions the interview actually got through.
| Candidate | Score | Recommendation | Communication clarity | Relevance of answers | Technical knowledge | Problem solving | Fit for the role | Confidence and presence | Behavioural fit | Completeness of answers | English language proficiency | Depth on prepared questions | Knockouts | Coverage | Date |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Anna Kowalska kowalska@example.com | 74 | Yes | 8.0 | 8.0 | 7.0 | 8.0 | 7.0 | 6.0 | 8.0★ | 7.0 | 9.0★ | 6.0 | ✓ 3 | 70 % | 15 Oct 2026 |
Lukas Berger berger@example.com | 72 | Yes | 7.0 | 8.0 | 8.0 | 7.0 | 8.0 | 7.0 | 6.0 | 6.0 | 8.0 | 6.0 | ✓ 3 | 74 % | 14 Oct 2026 |
Tomas Hruby hruby@example.com | 70 | Yes | 6.0 | 7.0 | 9.0★ | 8.0 | 7.0 | 6.0 | 5.0 | 7.0 | 6.0 | 8.0★ | ✓ 3 | 88 % | 13 Oct 2026 |
Marek Szabo szabo@example.com | 58 | Yes | 6.0 | 6.0 | 6.0 | 5.0 | 6.0 | 6.0 | 6.0 | 6.0 | 7.0 | 5.0 | ✓ 3 | 62 % | 16 Oct 2026 |
Elena Petrova petrova@example.com | 52 | Maybe | 7.0 | 6.0 | 5.0 | 5.0 | 5.0 | 6.0 | 7.0 | 5.0 | 8.0 | 4.0 | ✓ 3 | 55 % | 12 Oct 2026 |
Jakub Dvorak dvorak@example.com | 47 | Maybe | 5.0 | 6.0 | 5.0 | 4.0 | 5.0 | 5.0 | 5.0 | 4.0 | 6.0 | 4.0 | ✓ 3 | 48 % | 17 Oct 2026 |
Peter Weiss weiss@example.com | 44 | No | 6.0 | 5.0 | 6.0 | 5.0 | 4.0 | 5.0 | 4.0 | 4.0 | 7.0 | 5.0 | ✕ 1✓ 2 | 60 % | 13 Oct 2026 |
Ivana Horvat horvat@example.com | 38 | No | 5.0 | 5.0 | 3.0▼ | 4.0 | 4.0 | 4.0 | 6.0 | 3.0 | 7.0 | 2.0▼ | ✓ 3 | 35 % | 16 Oct 2026 |
Martin Lang lang@example.com | 31 | No | 4.0 | 4.0 | 3.0 | 3.0 | 3.0▼ | 4.0 | 4.0 | 3.0 | 5.0 | 2.0 | ✕ 1✓ 2 | 30 % | 12 Oct 2026 |
David Novak novak@example.com | 12 | No | 3.0 | 2.0▼ | 1.0▼ | 1.0▼ | 1.0▼ | 2.0 | 2.0 | 1.0▼ | 4.0 | 0.0▼ | ✕ 2✓ 1 | 10 % | 15 Oct 2026 |
The top three in one chart
Two to four candidates can be shown at once across all ten areas. Their totals are close, but their strengths are very different.
The top three side by side
Exactly what the hiring manager sees on the comparison screen, strengths, risks, the candidate's own words and how much of each prepared question they covered.
Anna Kowalska
- Clear, structured explanations of design decisions.
- Strong teamwork and mentoring examples.
- Less depth on JVM internals.
- Limited experience with high-traffic systems.
"I always write the API contract first and review it with the frontend team before any code."
Technical interview with a focus on performance and JVM tuning.
Lukas Berger
- Hands-on Spring Boot and microservices in production.
- Real experience with resilience patterns.
- No Kubernetes operations experience.
- Concurrency answers stay at definitions.
"Every payment request carries an idempotency key, so a retry never charges twice."
Technical interview with a Kubernetes troubleshooting scenario.
Tomas Hruby
- Deepest technical knowledge in the group, including concurrency.
- Covered almost every prepared topic.
- Short answers on teamwork and collaboration.
- English is functional but slower.
"count++ is read, add and write, so volatile alone does not make it safe, you need an AtomicInteger or a lock."
Meet the team to check collaboration and communication fit.
Why we compare this way
- •The same questions for everyone means the numbers can be compared. In free-form interviews you compare impressions, not people.
- •Nobody drops out because of one bad answer. The chart shows exactly where each candidate is strong and where they lose points.
- •A candidate with a lower total can still be the best fit for your team, which is why you get all the numbers, not just a ranking.
- •Candidates interviewed weeks apart can still be compared fairly.
Six things we check in every expert
Knockout criteria
Clear yes or no requirements, such as availability or certifications.
Must-have skills
Each with the level the project needs, from basic to expert.
Deep technical questions
Follow-ups that probe 3 to 6 knowledge areas per question.
Real scenarios
Case studies, system design, coding discussion and behavioural questions.
Language level
An optional assessment on the CEFR scale, from A1 to C2, for the working language.
Full transcript
Every answer in writing, and a recording when the candidate agrees to one.
What it means for your project
Consistent
Every candidate gets the same depth and the same rubric.
Evidence instead of CV claims
Every score points to what the candidate actually said.
Faster
Experts are interviewed in parallel and at any hour, so your team can start sooner.
Fair to candidates, too
Good experts have options. A respectful, transparent interview is part of why they choose to work with us.
Consent first
Candidates agree before the interview and know that their answers are processed by AI.
People make the decision
No decision is made by AI alone. A person reviews every recommendation, as GDPR Article 22 requires.
Room to say no
Candidates can skip a question. It is recorded as skipped, not held against them.
Questions about expert vetting
How do you vet experts?
Every expert completes an in-depth AI interview with AI Screenr, built from the project's requirements. A senior Jobrela specialist then reviews the scored report and the transcript and makes the final decision.
Does AI decide who gets the job?
No. The AI interviewer asks the questions and scores the answers against a rubric, but a person reviews every recommendation and makes the decision, as GDPR Article 22 requires.
What does the AI interview check?
Knockout criteria, must-have skills and their level, deep technical knowledge, real-world scenarios, communication and, if needed, the level of the working language on the CEFR scale.
How long does the interview take?
Between 15 and 60 minutes, depending on the role. Candidates do it whenever it suits them, in their own language.
What do clients receive after the interview?
A scored report with evidence quotes from the transcript, the knockout and must-have results, scores in ten areas and a suggested next step, plus a comparison of all candidates for the role.
Is the interview fair to candidates?
Every candidate gets the same questions and the same rubric, agrees to the interview beforehand and can skip a question without it counting against them.
Want experts who have proven what they know?
Tell us what your project needs, and we will introduce experts who passed the interview.