Interviews and your portfolio

A method for coding problems

A good doctor asks questions before writing a prescription. Interviewers want to see how you think, not only the answer.

1. Clarify. Restate the problem. Ask about input size, empty input, duplicates, negative numbers, ties, whether the input is sorted, what to return when there's no answer.
2. Examples. Work one or two small examples by hand, including an edge case.
3. Start simple. Say the brute-force idea and its cost, such as "check every pair, O(n²)".
4. Improve. Look for the data structure that removes the repeated work: a hash map, sorting, two pointers, a heap. Say the new cost.
5. Code clearly. Good names, small helpers. Talk as you write.
6. Test. Run your examples through the code by hand, then edge cases. Fix bugs calmly; finding your own bug is a good sign, not a bad one.

It's the same habit as professional work: understand the problem, check with examples, and test.

def best_gain(prices):
    """The most you could make buying once, then selling once later."""
    if len(prices) < 2:
        return 0
    lowest, best = prices[0], 0
    for price in prices[1:]:
        best = max(best, price - lowest)
        lowest = min(lowest, price)
    return best

cases = {"the example": [7, 1, 5, 3, 6, 4], "only falling": [9, 7, 5], "one price": [5], "empty": [], "all equal": [3, 3, 3]}
for name, prices in cases.items():
    print(f"{name:<12} {best_gain(prices)}")

Behavioral questions: STAR

Tell it like a short story: set the scene quickly, say what you did, and end with how it turned out. Answer "tell me about a time when..." with Situation, Task, Action, Result. Keep the situation short, spend most of the time on your actions, and end with a concrete result: "cut response time from 4 s to 600 ms", "the team adopted it". Prepare five or six stories: a hard bug, a conflict, a failure and what you learned, a project you led, a time you changed your mind.

import re

def has_result(line):
    return bool(re.search(r"\d", line)) and bool(re.search(r"\b(cut|raised|reduced|saved|grew|from|to)\b", line.lower()))

for line in ["Responsible for the checkout service",
             "Worked on improving API performance",
             "Cut p95 checkout latency from 2.1 s to 300 ms by adding a cache",
             "Reduced support tickets 30% with an agent that answers order questions"]:
    print("strong " if has_result(line) else "vague  ", line)

Keep growing

This field changes monthly. Doctors keep reading journals long after medical school, and you'll need the same habit. Read release notes and docs from the tools you use, read a paper a month, rebuild something to understand it, and teach what you learn. The habits from this academy, building, measuring and explaining, are what keep you professional as the tools change.

def pair_slow(nums, target):
    for i in range(len(nums)):
        for j in range(i + 1, len(nums)):
            if nums[i] + nums[j] == target:
                return (i, j)
    return None

def pair_fast(nums, target):
    seen = {}
    for j, n in enumerate(nums):
        if target - n in seen:
            return (seen[target - n], j)
        seen.setdefault(n, j)
    return None

nums = [8, 3, 11, 5, 7]
print(pair_slow(nums, 12), pair_fast(nums, 12))

seen = {}
for i, n in enumerate([4, 7, 1, 7]):
    seen.setdefault(n, i)
print(seen)

What gets people hired

Depth beats number of projects. One project that's deployed, tested and written up teaches an interviewer far more than ten tutorials. Be ready to explain why you made each decision, and what you'd change now.

Clarify before you code. Jumping straight in looks keen, and it often means solving the wrong problem. Two minutes of questions shows the judgment interviewers are looking for.

Think out loud. A math teacher wants to see your working, not only your answer. An interviewer can't give you credit for what they can't see. If you go silent for ten minutes and present an answer, they've learned little about how you work. If you say "I'm stuck on the empty case", they can help, and that conversation is part of what they're assessing.

Finish your stories with a result. "We improved the API" could mean anything. "Latency fell from 2 s to 300 ms" is a story they'll remember.