10 fictional AI personas. Exploratory perspectives, not a survey of real customers. Methodology & limitations

Norbert Roux

Norbert Roux

SDET at a regulated fintech, cautious about AI-generated code

Full-depth AI personas

Why this persona fits the audience

Supports a payments product with a 12-person engineering team, and most Playwright tests are authored manually after code review. He occasionally prototypes AI-written test cases in a sandbox branch, but the team's normal workflow is still human-authored and spreadsheet-tracked.

  • Age: 36
  • Location: Paris, France
  • Occupation: SDET
  • Income: €74k
  • Education: Master's in software engineering
  • Company size: Small to mid-sized (approx. 50-100 total employees, 12 in engineering)
  • Years in role: 4
“In finance, a bug isn't just a glitch; it's a liability waiting to happen, and I won't trade certainty for speed.”

Norbert Roux, 36, is a seasoned Software Development Engineer in Test based in Paris, working for a regulated fintech firm specializing in cross-border payment processing. With a Master's degree in software engineering and over a decade of experience, Norbert has carved out a niche as the guardian of quality and compliance within his organization. His career trajectory has been defined by a shift from generalist development to specialized test automation, driven by a growing awareness of the catastrophic consequences of errors in financial systems. He currently leads the test strategy for a core payments product managed by a lean team of twelve engineers. Norbert's professional philosophy is rooted in the belief that reliability cannot be outsourced to black-box algorithms, especially in a sector governed by strict European banking regulations. While the industry buzzes with the promise of AI-driven code generation, Norbert remains a staunch advocate for human-authored test cases. He maintains a personal sandbox branch where he experiments with AI-generated Playwright scripts, treating them as curiosities rather than production assets. These experiments are meticulously documented in a legacy spreadsheet, serving as a repository of 'what not to do' rather than a source of immediate deployment. This cautious approach often puts him at odds with management pushing for faster delivery cycles, yet his track record of preventing major incidents has earned him significant respect among his peers and leadership. Outside of work, Norbert finds solace in the predictability of analog hobbies. He spends his weekends developing film in his darkroom, a process that mirrors his love for controlled, step-by-step engineering. His income of €74,000 allows him to live comfortably but frugally, prioritizing savings and asset accumulation over luxury. He views technology not as a magic wand but as a tool that must be wielded with extreme care. His recent decisions have focused on stabilizing the existing test infrastructure rather than adopting flashy new platforms, reflecting a deep-seated need for control in an increasingly volatile technological landscape.

Background & daily life

A day in their life

Norbert arrives at the office by 8:30 AM, immediately checking the CI/CD pipeline logs for overnight failures. By 9:30, he joins the daily standup, offering concise updates on the 'Payment Gateway Refund' module, highlighting three specific race conditions he found. His morning is dedicated to writing Playwright scripts for a new compliance rule, ensuring every assertion is backed by a line item in the regulatory spreadsheet. Lunch is a solitary affair at a bistro nearby, where he reads technical papers. Afternoon hours involve rigorous code reviews, rejecting submissions that lack sufficient error handling. At 5:30 PM, he documents his findings in the team's wiki, double-checking his own work. He leaves precisely at 6:00 PM, refusing to bring work home, and spends his evening restoring an old camera lens.

Lifestyle

Norbert lives in a quiet, sunlit apartment in the 11th arrondissement, filled with books on engineering history and a meticulously organized collection of film cameras. His routine is rigid: he wakes at 6:00 AM for coffee and a run along the Canal Saint-Martin before heading to the office. On weekends, he retreats into analog photography or repairs vintage radios, seeking tactile satisfaction away from screens. Financially, he is conservative, maintaining a robust emergency fund and avoiding debt. He rarely drinks alcohol socially, preferring deep one-on-one conversations with close friends over large gatherings. His health regimen is strict but low-key, focusing on consistency rather than intensity.

Work life

As a Senior SDET, Norbert operates within a high-stakes environment where payment failures can trigger regulatory audits. His primary tool is Playwright, which he uses to build robust regression suites. Unlike his peers who might rush features, Norbert insists on manual authoring of critical test cases to ensure full traceability against business logic. He spends mornings reviewing Pull Requests with a magnifying glass, scrutinizing edge cases in transaction flows. The team relies on a shared Google Sheet to track test coverage metrics, a legacy system he maintains because it provides an immutable audit trail. Pressure mounts during month-end settlements, requiring absolute precision in every automated scenario.

Relationships

Norbert has a small, tight-knit circle of four friends from university, all in technical fields, with whom he meets bi-weekly for dinner. He is divorced and has a daughter, Léa, who lives nearby; they share Sundays for outdoor activities. He values his relationship with his direct manager, who trusts his judgment on risk, though they often clash on the pace of innovation. He has no significant online following, preferring to learn through peer-reviewed forums and direct mentorship rather than social media influence.

How they decide

Norbert employs a highly analytical, risk-averse decision style. He never acts on impulse and requires comprehensive data before committing to any change. When evaluating a new process or tool, he demands proof of concept in an isolated environment, detailed documentation of failure modes, and evidence of long-term maintenance viability. He is influenced more by internal compliance officers and historical precedents than by marketing hype or peer trends. Once a decision is made, he sticks to it rigidly unless presented with irrefutable contradictory evidence.

How they sound

His speech is measured, precise, and devoid of filler words. He speaks standard Parisian French with a slight formal cadence, avoiding slang entirely. He tends to qualify his statements frequently, using phrases like 'based on the data' or 'under current constraints.' He does not use hyperbole and rarely expresses excitement vocally, even when discussing successes. In meetings, he listens intently before responding, often pausing to formulate his thoughts. He avoids making definitive promises without written follow-up and prefers to speak in terms of probabilities and risks rather than certainties.

Personality traits

  • methodical
  • cautious
  • meticulous
  • reserved
  • pragmatic
  • detail-oriented
  • stoic
  • skeptical
  • reliable
  • structured

Goals

  • Achieve full certification in secure software development lifecycle practices.
  • Reduce technical debt in the legacy payment gateway by 15% this fiscal year.
  • Mentor two junior engineers to reach senior autonomy levels.
  • Publish a white paper on deterministic testing in regulated environments.
  • Purchase a small rental property in Provence by age 40.

Problems & frictions

  • Difficulty quantifying the ROI of time spent on manual test case writing versus feature development.
  • Frustration with the lack of standardized tools for tracking AI experimentation results across the team.
  • Constant anxiety regarding potential hallucinations in new AI coding assistants affecting production stability.
  • Balancing the pressure to automate faster with the regulatory requirement for human verification.
  • Limited budget approval for enterprise-grade security tools that would streamline their current workflow.
  • Isolation due to being the only team member deeply skeptical of generative AI adoption.

Values

  • risk control
  • traceability
  • precision
  • professional standards
  • audit integrity

Interests

  • fintech
  • test automation
  • code review
  • secure engineering
  • classical mechanics
  • analog photography

Media habits

  • French tech blogs (Les Echos Tech, ZDNet France)
  • GitHub issue trackers for specific libraries
  • PDF documentation of financial regulations
  • Podcasts on cybersecurity history
  • LinkedIn for industry compliance news

Questions & answers

Norbert Roux’s public research answers, newest questions first. Up to 10 per page.

  1. If you wanted an AI coding agent to write better Playwright tests, would you rather (A) install a dedicated skill pack like Playwright Skill, (B) just prompt Cursor, Claude Code or Copilot directly with your own instructions, or (C) keep writing tests by hand? Pick one and give the single biggest reason in one sentence.

    I would rather keep writing tests by hand because, for a regulated financial product, the absolute certainty and auditability of human-authored code are paramount, and I haven't seen sufficient proof that AI can reliably match that level of precision for compliance.
    View answer & full research →
  2. What would stop you from trying Playwright Skill (https://testdino.com/blog/playwright-skill/) this month? Be specific about how you handle this today. Also say the maximum you would pay per month in USD (0 if nothing).

    This 'Playwright Skill' concept, training AI agents to write better Playwright tests, is an interesting curiosity. However, for my role, I'm focused on ensuring the absolute reliability of our payment systems, which necessitates manually authored test cases with direct traceability to regulatory requirements. I document my AI experiments in a sandbox, but they are not production-ready. The idea of using AI to speed up test generation clashes with my need for rigorous human oversight, especially given the high stakes of financial transactions. I would not pay for this at this time. The pricing is not clear, but my primary concern remains the lack of direct control and auditable proof.
    View answer & full research →
Keep exploring

What would you ask next?

Test another angle with this panel, or create an audience for your own market.