Coding interview prep for international students in Germany

Sources: Wissenschaft weltoffen (DAAD and DZHW), Im Fokus, Internationale Studierende nach Herkunftsländern, read 2 November 2026; Wissenschaft weltoffen kompakt 2026, figures 6, 10, 11, 12 and 36, based on the Federal Statistical Office student statistics. We quote shares as published and do not combine them. We are not connected with the DAAD or the DZHW.

For a technical interview the subject matters less than it seems. Engineering, physics and economics students often write code in their degree, in Python, MATLAB or R, but rarely practise the classic interview topics: hashing, recursion, graphs and the cost of an algorithm as the input grows. Those topics can be learnt in a few months by anyone who already programs a little, and they are what this page is about.

import random
import time

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

def pair_fast(prices, budget):
    seen = {}
    for j, price in enumerate(prices):
        if budget - price in seen:
            return seen[budget - price], j
        seen[price] = j
    return None

print(pair_slow([12, 7, 30, 5, 18], 23), pair_fast([12, 7, 30, 5, 18], 23))

random.seed(2026)
for n in (1000, 2000, 4000):
    prices = [random.randrange(1, 10**6) * 2 for _ in range(n)]
    budget = 1                       # odd: no pair of even numbers can reach it
    start = time.perf_counter()
    pair_slow(prices, budget)
    slow = time.perf_counter() - start
    start = time.perf_counter()
    pair_fast(prices, budget)
    fast = time.perf_counter() - start
    print(f"n={n}: nested loops {slow:.3f} s, dictionary {fast:.5f} s")

Our own code and timings, measured with time.perf_counter, which the Python documentation describes as the clock with the highest available resolution for short durations. The budget is odd and every price even, so neither function ever finds a pair: that forces the worst case for both.

Read the nested-loop column as an interviewer would. Each time n doubles, the time roughly quadruples: 0.008, 0.031, 0.126. That is what quadratic means in practice, and it is far more convincing than reciting "O(n squared)". The dictionary column grows much more gently, because each price is looked at once and each lookup is, on average, constant time. The Python documentation gives key in d as O(1) on average and adds a warning worth repeating in an interview: in the worst case, when every key hashes to the same value, those operations take O(n) time.

Our mentors work in English every day with learners from many countries, and they correct the things that cost marks rather than accent or grammar: unclear variable names said aloud, long silences, and answers that start with code before the plan. Classes are live and in English only. Students who would like to practise the same answer in two ways, first slowly, then at interview speed, can ask for that in one-to-one sessions.

German time follows the Einheitenzeitgesetz: MEZ (UTC+1) in winter, MESZ (UTC+2) in summer, with the switch back on 25 October 2026 and forward on 28 March 2027. India keeps UTC+5:30. Sample times only.

Before an interview date, many students book two or three mock interviews in the final week. Tell the mentor the date and the language the company uses, and the sessions are planned backwards from it.