Skip to main content
Intertricks/2026

From Hours to Minutes: AI-Graded Mock Interviews with Expert Review and QR-Verifiable Results

From Hours to Minutes: AI-Graded Mock Interviews with Expert Review and QR-Verifiable Results

Intertricks needed a complete mock interview solution. Candidates attempt interviews, the attempts get graded, and the result lands back in the hands of qualified agency experts for final judgement. The problem was time: reviewing a single interview took hours, and because attempts could be submitted as video, there was no way to speed it up manually.

We built a platform that accepts both text and video interview attempts, processes and transcribes them using GPT, and produces a proposed grade with its reasoning attached. One of their experts then confirms or overrides that grade. Once approved, the platform issues the candidate's certification and publishes their interview result — and the result carries a QR code, so any third party can scan it and verify the outcome directly against the source record. The system was also built to take on other agencies, each managing their own students independently.

Review time dropped from hours to minutes. The expert's decision stayed exactly where it was.

Challenge

Intertricks is an agency that runs mock interviews for candidates preparing for real ones. Their own experts evaluated every attempt and issued the result. The model was sound — it just could not scale, and the reason was time.

Reviewing a single interview took hours. Every attempt had to be worked through in full, evaluated against criteria, and scored by hand. That put a hard ceiling on throughput: more candidates produced a longer queue, not more completed reviews. Turnaround became unpredictable for people who often needed a result by a deadline.

Video made it structurally worse. Interviews could be attempted by video, and there is no skimming a recording. An expert had to sit through the entire thing before forming any judgement. The format that captured the most signal was the one that consumed the most expert time.

They wanted to open the platform to other agencies. Intertricks did not just want their own operation running better — they wanted a system other agencies could work on too, each managing their own students independently, without anyone's data crossing over.

Results needed to be verifiable by outsiders. A mock interview result is only worth what someone else will trust. An employer or institution looking at a candidate's outcome needed a way to confirm it was genuine and unaltered — without emailing Intertricks and waiting for someone to check a record by hand.

The obvious fix — grade everything automatically — was not acceptable. The value of a mock interview sits in an expert's judgement of how the candidate actually performed. Removing that would have made the service fast and worthless. The speed had to come from somewhere else.

Solution

We built a system where AI absorbs the processing load and the expert keeps the judgement.

Text and video interview attempts. Candidates attempt interviews in either format. Video submissions run through a processing and transcription pipeline built on the latest GPT model, turning a recording into an accurate, structured text record. A long interview becomes something an expert can read, search and evaluate rather than sit through in real time.

AI grading with visible reasoning. Each processed attempt is evaluated against the interview criteria and returns a proposed grade with the reasoning attached — not just a score. That distinction is the whole design. The expert audits a judgement that has already been laid out, rather than reconstructing one from scratch.

The result goes back to a human expert. No certification is issued on an AI grade alone. A qualified expert sees the attempt, the proposed grade and the reasoning behind it, and confirms or overrides. Their role shifted from performing the review to verifying it — and that shift is where the hours-to-minutes gain comes from, without surrendering accountability.

Multi-agency student management. The platform was built so agencies beyond Intertricks can operate on it, each enrolling and managing their own students, tracking attempts and following progress in their own space — without any agency seeing another's data.

Automated certification issuance. Once an agency expert approves a grade, the platform generates and issues the candidate's certification automatically — no manual document preparation, no separate step between approval and delivery. Certification becomes a consequence of the review being signed off rather than a task someone has to remember.

QR-verifiable results. The interview result is a separate record from the certification, and it carries its own QR code. Scanning it resolves straight to the source record, so an employer, institution or partner can confirm the candidate's actual outcome is genuine and unaltered — in one scan, with no request to anyone and no waiting.

Built on Next.js, Node.js and PostgreSQL, with the GPT model handling video processing, transcription and grading.

How we worked. A paid one-week scoping phase settled the interview criteria structure, the grading logic, and how certifications and results would be issued and verified — before anything was built — decisions that are expensive to reverse once real data exists. From there, two-week sprints with weekly demos. The engineers who designed the pipeline are the ones who built it.

Results

Interview review time reduced from hours to minutes. The manual process required an expert to work through each attempt in full. With AI transcription and pre-grading, the expert verifies a structured result instead — the same decision, reached in a fraction of the time. This was Intertricks' largest operational constraint, and it is the one the platform removed.

Video interviews became practical at scale. Previously the slowest format to review, video attempts now pass through automated processing before reaching a human, removing the bottleneck that made them impractical to offer widely.

The platform can host other agencies, not just Intertricks. Each agency enrolls, tracks and manages its own students independently — turning an internal tool into something Intertricks can extend to others.

Certification issuance became automatic. Approved reviews produce certifications without manual document work, removing a step that previously sat between an expert's decision and the candidate receiving anything.

Results became independently verifiable. Each interview result carries a QR code resolving to the source record, so any third party can confirm a candidate's outcome is genuine — removing manual verification requests entirely.

Expert accountability preserved. Every result still passes a qualified human expert. Throughput increased without removing the person whose judgement gives the certification its value.

Ongoing engagement. The platform was built for extension. A central control panel giving Intertricks oversight across every agency on the system — student counts per agency, attempt status, and agency-by-agency comparison — is scoped as the next phase, along with payment integration.

Technologies & Tools

Next.jsNode.jsPostgreSQLOpenAI GPTDockerMySql

We hired TechnoBlick for a complete mock interview solution with AI grading and review, and then putting it back in the hands of agency experts. They did more than we were expecting and resolved our biggest problem — the time issue we were facing for reviewing the interviews. We also wanted the other agencies to be able to manage their students easily, and that was done pretty well.

CEO, Intertricks

Project Details

Year

2026

Duration

4

Budget

$under – $5K

Team Size

5 people

Live ProjectView →