Field notes · was Appendix A
The interview, from both sides
The strange ritual that stands between you and the work. It is a separate skill from engineering, learnable on its own terms — and being brilliant at the job does not make you good at the room where they decide whether to let you do it.
Last reviewed 2026-09-04
The gap, and why it is freeing#
The interview is not a measurement of how good an engineer you are. It is a measurement of how well you interview — a related but distinct game with its own rules, preparation and technique. That is the good news: you can get better at it deliberately, without becoming a different engineer.
You will sit on both sides of this table, often in the same year. Each side makes you better at the other, which is why this page is written twice.
Before: the refresh that is actually drillable#
A core of fundamentals gets probed again and again, and they are drillable. Not “do you know everything” — nobody does, and pretending to is its own red flag — but “have you refreshed the things this genre keeps asking about”. The daily work uses them through abstractions, so the raw forms go quiet.
Worth a weekend before a serious interview season
- Data structures, and the trade-offs between them — not the definitions, the reasons you would pick one.
- The design patterns you would actually reach for, and why.
- Object orientation and SOLID, stated cleanly, in your own words.
- Database fundamentals: indexing, normalisation, the shape of a good query, what makes a slow one slow.
- Dependency injection as a concept, not as your framework's configuration file.
- The genre's greatest hits — the recurring algorithm questions and classic system-design prompts. Work through them; do not memorise answers. Interviewers can smell rehearsal, and the point is understanding.
The technique for attacking the problem once it is on the board is its own discipline: the whiteboard method.
In the room#
Past a certain technical bar, the interview is won and lost on how you come across, how you talk about your past, and how you make the room feel about working with you. Every moment below is the same moment, seen from two chairs.
Why are you leaving?
Your side
Never say bad things about an old job or an old team. You think you are explaining why you left; what the room hears is how you will one day talk about them. Bitterness reads as a trait, not a circumstance. The honest-and-wise framing points forward: “my old job stopped being challenging, and I'm looking for somewhere I can grow.” Same truth, opposite valence.
The other side of the table
How someone talks about a former team is the single most predictive thing you will hear all hour, and it costs nothing to ask. You are not looking for people who have never had a bad manager. You are looking for whether the story is told with grievance or with perspective.
What do you know about us?
Your side
Specific, true enthusiasm for the actual work is disarmingly rare and disproportionately effective. Not flattery — one real thing you find genuinely interesting about what they are building. If you cannot find one, that is information about whether you want the job.
The other side of the table
Do not treat this as a loyalty test. Treat it as a signal about how the person prepares. And remember the traffic runs both ways: the strong candidate is reading you as hard as you are reading them, and the best people have options.
Where do you see yourself?
Your side
Frame a shared future, not a private ambition: a trajectory with them, growing into more responsibility and more depth. It answers the question they are actually asking — will this person stay and grow, or are we a stepping stone?
The other side of the table
You are trying to find out whether the role can hold this person for two or three years. Say what the path actually looks like at your company, including where it stops. Overselling a ladder that does not exist buys you an expensive twelve months.
The thing you do not know
Your side
Confidence is knowing what you are good at and owning it without flinching. Arrogance is needing to be better than the question. The difference is almost entirely in how you handle uncertainty: “I don't know that, but here's how I'd find out,” said with total ease, is confidence. Not being able to admit a gap is the other thing.
The other side of the table
“I don't know, here's how I'd find out” is a green flag and should be scored as one. A candidate who bluffs a technical answer will bluff a status update. Ask something outside their stack on purpose, and watch what they do with the edge of their knowledge.
Tell us your weakness
Your side
The fake weakness — “I work too hard,” “I care too much” — has been heard a thousand times and reads as evasion: you have just demonstrated you cannot be honest about yourself. A genuine limitation, owned plainly, plus the concrete thing you are doing about it, does triple duty: honest, self-aware, and a live demonstration that you treat faults as things to work on rather than hide.
The other side of the table
If you ask this, be worth answering honestly. A room that punishes a real answer trains everyone in it to give fake ones, and you have just spent your one chance at self-awareness on a ritual.
Why should we hire you?
Your side
Answer it as a fit, not a boast: here is what you need, here is what I bring, here is where they meet. Grounded in their actual situation. You are not arguing you are the best engineer alive; you are showing you are the right one for this.
The other side of the table
This question mostly measures composure. If you want signal about fit, describe the hardest real problem on your team right now and ask how they would start.
Are you a team player?
Your side
Asserted, it is worthless. Shown, it is everything. Have real stories ready: the time you unblocked someone, mentored a junior, took the unglamorous work, handled a disagreement well. Underneath every technical question the room is asking one thing — do I want this person on the team?
The other side of the table
The team that will live with this person should meet them, and their read should carry real weight. But keep the question concrete — can we work with this person, would we trust them at the table — and never let “fit” quietly come to mean “familiar”. That distinction gets a whole chapter in the book, and it deserves one.
The tactical rules#
- Do not anchor yourself on salary
- Naming a number first, early, is a position you can only lose from: too low and you have capped yourself, too high and you have created friction before they have decided they want you. Deflect gracefully — “I'd like to understand the role first; I'm sure we can find a number that works if we're a fit” — and let the value-establishing happen before the number-naming.
- Elaborate — never answer in one word
- One-word answers starve the conversation, make the interviewer do all the work, and hide the only thing being measured: how you think. Give the answer, then the reasoning, then the example.
- Ask real questions
- “No, I think you covered everything” is a quietly fatal answer. Good questions show you are already thinking like someone in the role — and they are your own diligence. You are interviewing them too.
- Dress a notch above the room
- Unfashionable and I will defend it. Not because clothes measure competence, but because effort is a cheap signal of respect, and cheap signals of respect are worth sending. You can dress down once you know the culture.
After the interview, and before the next one#
- Follow up. A short, specific note — thanking them, referencing something real from the conversation, reaffirming interest. Rare enough that it stands out, easy enough that skipping it is just lost ground. You have the connection warm; press it gently before it cools.
- Keep notes, and run the search as a project. Who, what you learned, what to follow up. You will interview better in the tenth conversation because you captured the lessons of the first nine.
- Have a two-year plan and a way to measure it. It answers “where do you see yourself” for free, and it serves you far more outside the room than in it. The job is a step on a path; know the path.
- Do not wait for the perfect time. It does not come. The doors worth walking through are rarely the ones that opened when conditions were ideal — the only way to catch one is to already be the kind of person who is ready.
The question bank#
Eight that come up almost every time, and what a strong answer is doing.
Tell me about yourself.
Ninety seconds, shaped as a trajectory rather than a CV reading: where you started, the turn that changed how you work, what you are looking for next. Land on the reason you are in this room.
Walk me through a project you are proud of.
Pick one where you can name the constraint, the trade-off you chose, and the thing you would do differently. Pride plus a named regret reads as seniority; pride alone reads as marketing.
Tell me about a time you disagreed with someone.
The good version has the disagreement happening early and in public, about the work, and resolved without anyone being humiliated. If your story ends with you being proved right, add what it cost.
How do you handle a tight deadline?
The honest answer includes what you cut and who you told. “We worked weekends” is not a method; “we shipped the smallest thing that answered the question, and wrote the loan down” is.
What is your biggest failure?
Choose a real one with a real cost, tell it without protagonism, and say what changed in how you work. A failure with no behavioural consequence is an anecdote, not an answer.
How do you keep learning?
Specifics beat philosophy: what you are learning right now, how you make the hours, one thing you have shipped or broken as a result.
Design X for me.
Ask what it is for and who uses it before you draw anything. The strongest signal you can send in a system-design round is refusing to solve a problem you have not defined.
Do you have any questions for us?
Ask what the team argues about; what happens when a release goes wrong; who decides when two teams disagree; what someone in this role got wrong in their first six months. Those four questions tell you more about a company than any careers page.
The one-page card#
For the night before. Print this page — everything but this list is hidden.
- Refresh the fundamentals; work three of the genre's greatest hits.
- Find one specific, true thing you like about what they build.
- Prepare the leaving answer, pointed forward, with no bitterness in it.
- Prepare one real weakness and the concrete thing you are doing about it.
- Prepare three stories: unblocked someone, took the boring work, disagreed well.
- Decide in advance to deflect the salary question until they want you.
- Write down four questions to ask them. Not three.
- Say “I don't know, here's how I'd find out” at least once, comfortably.
- Elaborate. Reasoning, then example. Never one word.
- Send the follow-up the same day.
The interview is a flawed instrument measuring an approximation of a related skill, on both sides. Knowing that — holding it lightly, preparing for it deliberately, and never mistaking it for the work itself — is most of getting good at it.