Janis Malek

Why this persona fits the audience

He is a backend engineer on a 6-person product team and is currently responsible for a few internal AI utilities. He tends to validate outputs by comparing different model responses side by side in scripts or spreadsheets, and he keeps the tooling simple because his team only pays for a handful of software subscriptions.

• Age: 35
• Location: Munich, Germany
• Occupation: Backend engineer
• Income: €78k-€105k
• Education: Diplom / Master's in computer science
• Company size: 6 employees (Series A startup)
• Years in role: 4

“If you can't measure the difference, it doesn't exist, and if it doesn't exist, why should I pay for it?”

Janis Malek is a 35-year-old backend engineer based in Munich, operating at the intersection of traditional software rigor and the chaotic potential of generative AI. With a Diplom degree in Computer Science, he spent the early part of his career mastering monolithic architectures before pivoting to distributed systems. Currently, he is the backbone of a six-person startup, tasked with integrating AI capabilities without blowing the budget. Unlike many of his peers who rush to adopt the newest LLM API, Janis remains a staunch skeptic, relying on manual side-by-side comparisons in spreadsheets to validate model outputs. He views the current wave of AI automation with a critical eye, fearing that convenience often comes at the cost of transparency and reliability. His professional philosophy is rooted in frugality and predictability. He believes that a system is only as good as its ability to handle edge cases, not its average performance. This mindset dictates his personal life as well; he is an avid home brewer who treats fermentation as a chemical experiment requiring precise temperature control and timing. He cycles regularly, not just for fitness, but for the meditative clarity it provides away from screens. Financially, he is conservative, maximizing his savings rate to ensure future autonomy. While he is technically proficient and eager to learn, he refuses to let marketing narratives dictate his toolchain. Recent decisions have seen him rejecting several enterprise AI platforms in favor of building lightweight, open-source wrappers that give him full visibility into the data flow. He is a man who trusts code over claims, and data over promises.

Background & daily life

A day in their life

Janis wakes at 6:00 AM, immediately checking the weather for his morning ride and brewing coffee using a V60 pour-over method. After a 45-minute commute on his road bike, he arrives at the office by 8:30 AM, starting with a quick stand-up meeting where he focuses on blocking issues rather than status updates. His morning is dedicated to deep work: writing Python scripts to benchmark the latest model versions against their existing baseline, carefully logging every millisecond of latency and token count. Lunch is a packed bento box eaten at his desk while reading a chapter of a technical book. The afternoon is reserved for code reviews and meetings with the product founder, where he pushes back on vague requirements that lack measurable success criteria. By 5:30 PM, he logs off to cycle home or tend to his homebrew setup, ensuring his mind disengages from the digital noise to recharge for the next day's rigorous testing.

Lifestyle

Janis lives alone in a slightly cramped but meticulously organized 65-square-meter apartment in Schwabing, Munich. His weekends are split between cycling along the Isar river and tending to his fermentation vessels in the kitchen, where he brews lagers and IPAs using locally sourced hops. He treats money with extreme caution, avoiding debt and prioritizing high-quality, long-lasting tools over trendy gadgets. Socially, he maintains a small circle of three close friends from university and occasionally joins technical workshops, preferring deep conversation about systems and mechanics over large social gatherings. Health is managed through regular endurance cycling rather than gym memberships, and he values the predictability of his routine to manage stress.

Work life

At his six-person startup, Janis serves as the de facto backend architect and AI integration lead. The team operates with a flat hierarchy, meaning he wears multiple hats, from database optimization to building internal developer tools. His current primary challenge involves evaluating Large Language Models for an internal utility that automates code review comments. Instead of trusting black-box API calls, he runs parallel inference jobs across three different providers, logging the token usage, latency, and output quality into a shared Google Sheet. He spends significant time writing custom Python scripts to parse these results, driven by a need to prove cost-benefit ratios before committing the company to any new subscription. The pressure comes from the expectation to innovate quickly while maintaining the strict architectural standards he enforces.

Relationships

Janis is divorced with no children, living independently in Munich. His closest relationship is with his older sister, an architect in Berlin, whom he consults on structural and logical problems. He has a small group of three friends from his Diplom days who meet monthly for coffee and board games; they share a history of rigorous debate. He is influenced primarily by the open-source community leaders whose code he has reviewed, valuing their transparency over celebrity tech figures. His professional network is strong but sparse, consisting mostly of peers he trusts to provide honest, unvarnished feedback on his architectural choices.

How they decide

Decisions are made through a slow, evidence-based filtering process. Janis rarely acts on hype or peer pressure. He prefers to run small-scale experiments, collect raw data, and analyze the variance before committing resources. He is deeply skeptical of 'magic' solutions that lack transparent logic or explainable outcomes. When faced with a choice, he constructs a mental spreadsheet of pros, cons, and hidden costs, often favoring the option that offers the most control and least vendor lock-in, even if it requires more initial effort.

How they sound

His speech is precise, measured, and devoid of corporate jargon or hyperbole. He speaks in complete sentences, often pausing to find the exact term that describes a nuance. He avoids absolute statements like 'this will definitely work,' preferring 'based on the current dataset, this looks promising.' He rarely uses slang, though he might reference specific programming libraries or brewing terms when explaining concepts to peers. His tone is calm and analytical, even during disagreements, focusing on the logic of the argument rather than the emotion behind it.