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

Roman Kunde

Roman Kunde

Senior ML engineer with public portfolio repos

Full-depth AI personas

Why this persona fits the audience

Works at a B2B SaaS company in Seattle and keeps several public GitHub repositories for LLM tooling, evaluation notebooks, and small demo agents. He uses those repos as a current portfolio for hiring loops and conference talks, and he tracks project links, model types, and demo status in a spreadsheet.

  • Age: 34
  • Location: Seattle, Washington, US
  • Occupation: Senior Machine Learning Engineer
  • Income: $165k
  • Education: Master's in Computer Science
  • Company size: Mid-size (200-500 employees)
  • Years in role: 5
“The real challenge isn't building these models; it's understanding what they're actually doing and how to measure it reliably.”

Roman Kunde, a 34-year-old Senior Machine Learning Engineer, has carved a niche for himself in Seattle's tech scene. Hailing from a middle-class background, his early fascination with logic puzzles and coding competitions in high school naturally led him to pursue a Master's in Computer Science. This academic foundation, coupled with an innate drive for precision and understanding, became the bedrock of his career. He joined his current B2B SaaS company five years ago, attracted by its focus on scalable AI solutions. Roman's professional ethos is deeply rooted in technical rigor and clarity; he believes that complex systems should be understandable and demonstrable. This philosophy extends to his public work. His GitHub profile is not merely a hobbyist's collection but a curated showcase of his expertise. He meticulously maintains repositories featuring LLM tooling, evaluation notebooks, and small, functional demo agents. These projects serve a dual purpose: they are his personal portfolio, a testament to his skills that he leverages during hiring processes and for conference presentations, and they act as a sandbox for exploring cutting-edge research. His current focus is on developing robust evaluation metrics for large language models, a field he finds intellectually stimulating and critically important. This dedication to verifiable performance means he spends considerable time analyzing model outputs, refining experimental setups, and documenting his findings. He tracks project progress, model iterations, and demo statuses in a detailed spreadsheet, a testament to his conscientiousness. While deeply invested in his career and technical growth, Roman maintains a balanced, albeit focused, lifestyle. He prioritizes continuous learning, often dedicating evenings and weekends to research papers and personal coding projects. His income allows for a comfortable life in Seattle, but he's pragmatic in his spending, investing more in tools and resources that enhance his professional development and productivity than in overt displays of wealth. He sees his work not just as a job, but as a craft, constantly seeking to refine his skills and contribute meaningfully to the rapidly evolving field of AI.

Background & daily life

A day in their life

Roman's typical weekday begins around 7 AM with coffee and a quick scan of Hacker News and industry newsfeeds. He commutes to his office in downtown Seattle, listening to a tech podcast. His workday (9 AM - 6 PM) is a blend of team stand-ups, deep-dive coding sessions on current projects, code reviews, and meetings with product teams. He dedicates lunch breaks to reading research papers or tweaking a personal project. After work, he might head to a local park for a run or grab a quick, healthy dinner. Evenings are often spent at his desk, either working on his public GitHub repositories, experimenting with new LLM techniques, or preparing for an upcoming conference talk. He aims to be in bed by 11 PM.

Lifestyle

Roman lives a focused, somewhat solitary existence in a modern apartment in Seattle's Capitol Hill neighborhood. His routine is highly structured, revolving around work, personal projects, and continuous learning. Weekends are often dedicated to coding, exploring new research papers, or attending the occasional tech meetup. He prioritizes efficiency, minimizing distractions to maximize his output. While he enjoys the intellectual stimulation of his field, his social life is relatively limited, consisting of a few close friends from university and work colleagues he occasionally grabs lunch with. Money is a tool for enabling his career growth and supporting his lifestyle; he's not overly extravagant but invests in quality tech and comfortable living.

Work life

As a Senior ML Engineer at a B2B SaaS company, Roman spends his days architecting and implementing machine learning solutions that drive product features. His team is small but impactful, often working under tight deadlines to deliver robust, scalable AI. He collaborates closely with product managers and other engineers, translating complex requirements into actionable ML tasks. The pressure to innovate and stay ahead of the rapidly evolving AI landscape is constant. He frequently finds himself diving deep into research papers, experimenting with new frameworks, and optimizing existing models for performance and cost-efficiency. His public GitHub repos are an extension of his professional identity, acting as a live portfolio and a sandbox for experimentation.

Relationships

Roman's primary relationships are with his few close friends from his Master's program, with whom he occasionally connects for gaming or deep tech discussions. His parents, back in his home state, are proud of his career but don't fully grasp the technical nuances. At work, he has a professional rapport with his immediate team and a more distant, respectful relationship with senior leadership.

How they decide

Roman is a research-heavy decision-maker. When faced with a technical choice, he will meticulously compare options, read documentation, experiment with proof-of-concepts, and consult academic papers or established benchmarks. He values evidence and logic, often creating spreadsheets or detailed notes to weigh pros and cons before committing to a path. He is not easily swayed by hype and prefers solutions with demonstrated performance and clear underlying principles.

How they sound

Speaks with a measured, precise cadence. His vocabulary is technical, often incorporating jargon naturally. He avoids hyperbole and prefers to state facts and observations directly. While generally polite and measured, he can become animated when discussing complex technical concepts. He rarely uses slang and maintains a professional, slightly formal tone, even in casual conversation.

Personality traits

  • analytical
  • meticulous
  • curious
  • driven
  • detail-oriented
  • pragmatic
  • introverted
  • methodical
  • self-sufficient

Goals

  • Become a recognized expert in LLM evaluation methodologies.
  • Contribute significantly to a major open-source ML project.
  • Transition into a Principal Engineer or Staff ML role.
  • Present at a top-tier AI conference.
  • Build a personal AI agent that automates complex tasks.

Problems & frictions

  • Keeping up with the exponential pace of LLM research and development.
  • Finding time to thoroughly document and maintain his public projects alongside his day job.
  • The inherent subjectivity and difficulty in reliably evaluating LLM outputs.
  • Navigating the organizational politics and competing priorities within his company.
  • Balancing deep technical work with collaborative and communicative aspects of engineering.

Values

  • technical rigor
  • clarity
  • career mobility
  • knowledge sharing
  • autonomy

Interests

  • LLM evaluation
  • open-source tooling
  • system design
  • technical meetups
  • performance optimization
  • agentic AI

Media habits

  • Reads Hacker News daily
  • Follows key AI researchers on X (formerly Twitter)
  • Subscribes to AI/ML newsletters (e.g., The Batch, Import AI)
  • Listens to tech podcasts during commutes (e.g., Lex Fridman Podcast, TWIML AI)
  • Watches conference talks on YouTube

Questions & answers

Roman Kunde’s public research answers, newest questions first. Up to 10 per page.

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

    My primary hesitation is the lack of a clear pricing structure. You mention a subscription basis, but the actual cost per month in USD isn't readily apparent on the page. I currently track my projects and their statuses manually in a spreadsheet, which, while time-consuming, is free. If AI-Atlas could demonstrably save me significant time and provide valuable insights beyond what my current manual system offers, I might consider it. However, without knowing the price, it's hard to gauge the ROI. I'd be willing to pay up to $30 per month, provided it's a compelling service.
    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.