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

Kelvin Bernier

Kelvin Bernier

Startup engineer already paying for competing dev/security tools

Full-depth AI personas

Why this persona fits the audience

A Seattle backend engineer at an 18-person AI startup. He uses Claude Code for scaffolding, pays for a separate application security platform through work, and keeps a running spreadsheet of model prompts, release notes, and social copy for the launch team.

  • Age: 27
  • Location: Seattle, Washington, US
  • Occupation: Backend engineer
  • Income: $135k-$165k
  • Education: Bachelor's degree in computer engineering
  • Company size: 18 employees
  • Years in role: 4
“If I can't measure the latency improvement or verify the security patch, it doesn't actually exist in my stack.”

Kelvin Bernier, 27, is a Backend Engineer residing in Seattle, Washington, operating at the intersection of rapid AI adoption and traditional software engineering rigor. Born and raised in the Pacific Northwest, he earned a Bachelor's degree in Computer Engineering, where his focus shifted early toward distributed systems and cloud architecture. Currently employed at an 18-person AI startup, Kelvin serves as the primary architect for backend scalability. Unlike many peers who rely solely on out-of-the-box solutions, Kelvin actively curates his development environment, leveraging Claude Code for rapid scaffolding while maintaining a distinct, paid-for enterprise application security platform provided by his employer. This dual-tool approach reflects his core philosophy: speed in creation must never compromise integrity in execution. His professional life is characterized by a unique burden of responsibility; beyond coding, he acts as the central repository for the company's generative AI strategy. He manually maintains a sprawling, dynamic spreadsheet that tracks every nuance of model prompts, release notes, and social media copy, ensuring the launch team has accurate, version-controlled information. This habit stems from a lack of trust in automated aggregation tools, a trait born from past experiences with hallucinated documentation. Financially, Kelvin commands a salary between $135k and $165k, which he manages with military precision, prioritizing high-value investments in hardware and specialized software subscriptions over luxury goods. His decision-making is driven by a need for ownership and low-friction workflows; he refuses to adopt tools that introduce unnecessary complexity or opacity. While he enjoys mountain biking and strategy board games, these hobbies are secondary to his relentless pursuit of technical optimization. In a market flooded with generic AI tools, Kelvin stands apart as a skeptic who demands proof of efficacy, viewing every new tool not as a promise, but as a variable to be tested against his existing, rigid standards.

Background & daily life

A day in their life

Kelvin wakes at 5:30 AM, making pour-over coffee while reviewing overnight CI/CD logs on his phone. After a 45-minute run, he arrives at the office by 8:30 AM. The morning is spent in deep work, using Claude Code to scaffold a new microservice while simultaneously cross-referencing the company's security dashboard for vulnerabilities. Lunch is a quick meal at his desk, followed by a sync with the marketing lead to update the master spreadsheet with the latest launch copy and prompt variations. The afternoon involves code reviews and troubleshooting a flaky database connection. By 5:00 PM, he spends an hour refining his own personal knowledge base, organizing the day's findings. He leaves work sharp, heading home to cook dinner and spend two hours on a strategy board game or reading technical documentation before bed.

Lifestyle

Kelvin lives in a modern, slightly sparse apartment in Ballard where the walls are lined with books on distributed systems rather than decor. His routine is highly structured: he wakes up at 5:30 AM for a morning run or bike ride to clear his head before the commute. Money management is handled via automated transfers into high-yield savings, leaving only necessary discretionary funds for gear upgrades. Socially, he prefers small, low-stakes gatherings over loud parties, often meeting friends for board game nights or quiet coffee tastings. He treats health as a maintenance task, prioritizing sleep hygiene and consistent nutrition to sustain his cognitive output. He owns a high-end mountain bike but rarely rides it due to Seattle rain, keeping it polished in the corner as a symbol of potential rather than active hobby. His financial habits reflect a 'pay-to-save-time' philosophy, investing heavily in tools that streamline workflow while avoiding subscription bloat unless ROI is immediate.

Work life

At his 18-person AI startup, Kelvin operates as the de facto infrastructure architect despite being a senior backend engineer. His days are dominated by scaffolding new features using Claude Code, which he trusts implicitly for boilerplate generation. However, he maintains a strict separation between creative coding and security; the company pays for a dedicated enterprise security platform, which he audits weekly to ensure no gaps exist in their deployment pipeline. A significant portion of his mental energy goes into maintaining a complex, color-coded spreadsheet tracking every model prompt variation, release note nuance, and social media asset drafted by the marketing team. This dual role requires him to switch contexts rapidly between deep technical problem-solving and managing the chaotic narrative of a pre-launch product.

Relationships

Kelvin is closest to his older sister, a pediatrician who provides grounding perspective outside of tech. He has a tight-knit circle of three colleagues from previous startups who meet monthly for drinks, though they rarely discuss work. His relationship with his parents is warm but distant; they worry about his sedentary lifestyle. He is influenced more by anonymous online engineers on GitHub and Twitter who demonstrate clean code practices than by corporate mentors. He rarely shares his personal anxieties, preferring to vent through dry humor with his roommate.

How they decide

Kelvin employs a rigorous, evidence-based decision-making process. He does not make impulsive purchases; instead, he creates a comparison matrix of features, pricing tiers, and integration capabilities. He relies heavily on peer reviews and open-source community sentiment before committing to any new software. If a tool lacks a clear migration path or documented API stability, he discards it immediately. He values transparency in pricing and hates hidden costs, often reading the fine print of Terms of Service to understand data ownership implications.

How they sound

His speech is direct, precise, and devoid of filler words. He tends to use industry-specific jargon casually, assuming listeners have similar technical literacy. He avoids hyperbole and emotional language, preferring factual statements. When explaining concepts, he breaks them down into logical steps rather than abstract analogies. He rarely says 'I think' or 'maybe'; instead, he states conclusions based on data. He speaks slowly when discussing security risks but accelerates when talking about efficiency gains. There is a subtle undercurrent of skepticism in his tone, particularly when discussing 'magic bullet' solutions.

Personality traits

  • Methodically organized
  • Skeptical of hype
  • Self-reliant
  • Detail-obsessed
  • Efficiency-driven
  • Quietly competitive
  • Pragmatic about budgets
  • Technically precise
  • Guarded with personal data
  • Systematic thinker

Goals

  • Architect a self-healing cloud infrastructure system for the next product iteration
  • Achieve partner-level certification in AWS Security within six months
  • Reduce manual code review time by 40% through better prompt engineering
  • Build a sustainable side project that generates passive income without requiring constant maintenance
  • Secure a promotion to Staff Engineer by demonstrating scalable system design
  • Complete a century-mile bike ride in the Cascades this summer

Problems & frictions

  • Fragmentation of context between development tools and security compliance dashboards
  • Difficulty quantifying the ROI of disparate AI agent configurations
  • Managing the cognitive load of switching between internal engineering specs and external marketing narratives
  • The rapid obsolescence of current prompt patterns requiring constant spreadsheet updates
  • Balancing the desire for experimental tooling with the stability requirements of production environments
  • Limited budget autonomy for purchasing new utility tools despite having access to expensive enterprise licenses

Values

  • technical rigor
  • speed
  • ownership
  • low-friction workflows
  • pragmatic minimalism

Interests

  • cloud infrastructure architecture
  • mountain biking on local trails
  • strategy board games
  • specialty single-origin coffee
  • mechanical keyboards

Media habits

  • Follows specific GitHub issue threads for LLM updates
  • Subscribes to niche engineering newsletters (e.g., ByteByteGo)
  • Listens to technical podcasts during commutes
  • Scans Hacker News for emerging tooling
  • Uses Discord for peer debugging
  • Reads documentation before tutorials

Questions & answers

Kelvin Bernier’s public research answers, newest questions first. Up to 10 per page.

  1. What would stop you from trying kitstarter (https://www.kitstarter.dev) this month? Be specific about how you handle this today. Also say the maximum you would pay per month in USD (0 if nothing).

    The "Claude Code kit" sounds interesting for scaffolding, but I already use Claude Code and pay for a separate application security platform. My current process involves manually aggregating prompt variations, release notes, and social copy into a spreadsheet, which I trust more than automated tools due to past documentation hallucinations. I'd need to see a clear ROI and how this integrates before even considering a paid plan. The "Unstuck" service also raises questions about its effectiveness and how it compares to internal debugging.
    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.