Technical Decisions and Tradeoffs

Topics Covered

What Decision Questions Test

Why decision questions exist

What the rubric looks for

What interviewers are worried about

Judgment is the real topic

Context makes the decision

Variants of the question

How decisions scale with level

Reversible and irreversible decisions

Decisions are made by people

Decisions include operational cost

What a strong answer sounds like

Decision stories and technical depth

Preparing the details

Pick decisions you personally drove

Avoid decisions that were obvious

Decisions leave artifacts

Decisions with numbers behind them

How many decision stories to prepare

Tell me about a difficult technical decision you made. Describe a trade off you had to make between speed and quality. Tell me about a technical choice you would make differently today. Technical decision questions sit between behavioral and technical interviews: the interviewer wants to see your judgment, not just your knowledge. This lesson explains what interviewers look for, how to choose and structure a decision story, how to explain trade offs clearly, how to talk about decisions that went wrong and how to handle the hardest variants.

Why decision questions exist

Most engineering work involves choices between imperfect options: build or buy, one database or another, ship now or harden first, a quick fix or a proper redesign. Interviewers cannot watch you make these choices on the job, so they ask you to describe one. The story reveals how you think: what you considered, who you involved, how you weighed risks and how you lived with the consequences. These are signals no coding exercise can capture.

What the rubric looks for

Decision rubrics usually score several behaviors. Did the candidate frame the decision clearly, including the constraints? Did they consider real alternatives, not straw men? Did they weigh trade offs explicitly, with criteria that mattered for the situation? Did they involve the right people and gather evidence? Did they make the decision at the right time and communicate it? And did they reflect honestly on the outcome? Each is something the interviewer will try to record.

What interviewers are worried about

Interviewers also screen for risks. They worry about candidates who choose technology because it is new or interesting, who cannot explain why alternatives were rejected, who make large irreversible decisions alone, who ignore operational cost, or who defend every past decision as correct. At senior levels, they also worry about candidates who over engineer, building for scale that never arrives, or who never revisit decisions as circumstances change.

Judgment is the real topic

Although these questions involve technology, the real topic is judgment. Interviewers are less interested in whether you chose the right database than in how you chose it. A candidate who chose a less popular option for well reasoned, context specific reasons scores better than one who chose the popular option because everyone else did. Keep the focus on reasoning, not on proving that your choice was objectively best.

Context makes the decision

Good decisions depend on context: team size, deadlines, existing systems, expected growth, budget, skills and risk tolerance. The same choice can be right in one company and wrong in another. Strong stories make the context explicit early, so the interviewer can judge the decision on its merits. A team of four with a deadline in six weeks should make different choices from a platform team of forty.

Variants of the question

Decision questions take many forms: a difficult technical decision, a trade off between speed and quality, a time you chose not to use a technology, a decision you made with incomplete information, a time you reversed a decision, a decision you disagreed with, a time you chose simplicity over flexibility. Two or three well prepared decision stories, with different kinds of trade offs, cover most of these.

How decisions scale with level

Junior engineers make decisions within a task, such as choosing a data structure or library. Mid-level engineers make decisions within a feature or component, such as an API design. Senior engineers make decisions for a service or team, such as storage choices or service boundaries. Staff engineers make decisions that affect many teams for years, such as platform standards or architectural direction. Choose stories at the scope you are targeting.

Interviewers scored framing, real alternatives, explicit trade offs and reflection, and expected more care for decisions that were hard to undo.

Reversible and irreversible decisions

A useful distinction is between decisions that are easy to reverse, such as a library choice behind an interface, and those that are hard to reverse, such as a primary database or a public API. Strong engineers move quickly on reversible decisions and take more care on irreversible ones. Mentioning this distinction in your story, and showing that your process matched the stakes, is a strong judgment signal.

Decisions are made by people

Technical decisions are also social. Someone must agree, someone must build it and someone must operate it. Strong stories show how the candidate involved the people affected, such as the on call engineers who would run the system, or the team who would migrate to it. Decisions made in isolation often fail in adoption, even when they are technically sound.

Decisions include operational cost

Many candidates focus on build time and ignore operational cost: who will run this, how it will be monitored, what happens when it fails at three in the morning. Interviewers, especially at senior levels, look for awareness of operational cost. A decision story that mentions on call load, upgrade burden or monitoring needs shows that you think beyond the launch.

What a strong answer sounds like

A strong decision answer is structured and honest. The candidate explains the context and constraints, names two or three real options, states the criteria that mattered most, explains how each option scored, describes who was involved and how the decision was made, gives the outcome and reflects on what they would do differently. It avoids technology boasting, straw man alternatives and defensiveness.

Decision stories and technical depth

Decision stories invite follow up technical questions: why not use a queue there, how did you handle consistency, what was the failure mode. Be ready to go one or two levels deeper on the technical details, but keep returning to the reasoning. The interviewer is testing whether your judgment is grounded in real understanding, not whether you can recite documentation.

Preparing the details

Decision stories depend on specifics: the constraints, the options considered, the criteria, the numbers that informed the choice, the people involved, the outcome and what you learned. Write these down before the interview. Many candidates remember the option they chose but forget why they rejected the others, which is often what interviewers probe most. Reconstruct the alternatives from design documents if you can.

Pick decisions you personally drove

Choose decisions where you were the main decision maker or the person who framed the options and made the recommendation. Stories about decisions your lead made, which you implemented, are weaker. If the decision was made by a group, be clear about your role: I wrote the comparison and recommended the option we chose, and the tech leads made the final call.

Avoid decisions that were obvious

A decision that any engineer would make the same way is not a good story. Choose decisions with real tension, where reasonable people could disagree and the options had genuine strengths. If the interviewer could not imagine choosing differently, the story will not reveal your judgment. Look for decisions where you rejected something attractive for a good reason.

Decisions leave artifacts

Good decisions usually leave a trail: a design document, a comparison table, a prototype branch, a meeting note, an incident review. Before the interview, look for these artifacts and reread them. They remind you of the options you rejected and the numbers that guided you, which are the details that make a story credible. They also stop you from rewriting history in a way that makes the decision look cleaner than it was.

Decisions with numbers behind them

Interviewers trust decisions that were grounded in evidence. If you ran a benchmark, a load test, a cost estimate or a small user study before deciding, mention it. Even a rough estimate counts: we expected about two hundred writes a second at peak, which a single database could handle for at least two years. Evidence shows that you did not decide on instinct alone, and it gives the interviewer something concrete to probe.

How many decision stories to prepare

Prepare two or three decision stories with different shapes: one where you chose speed, one where you chose long term health and one where the decision went wrong. Together they cover most variants. Each story should be able to stand alone, but having several lets you choose the one that best fits the exact question, rather than forcing a single story to answer every prompt.

Judgment, not technology

Decision questions test how you frame a choice, consider real alternatives, weigh trade offs against context specific criteria, involve the right people, match care to reversibility, account for operational cost and reflect honestly. Interviewers are wary of technology for its own sake, straw man alternatives and defending every past choice.