Math problems are arithmetic in a costume

Math and geometry problems in interviews are small number-theory tricks dressed up as questions. Modular arithmetic, GCD, fast exponentiation, sieves, squared distance — that's most of it. The interview tests whether you recognize which trick the problem wants.

Math problems in interviews aren't calculus exams. They test a small set of number theory tricks and geometric reasoning you can learn in an afternoon. The key is recognizing which trick applies.

Diagram

Copy source

Rendering diagram...

The Tricks You Need

1. Modular Arithmetic

Most "math" problems reduce to modular arithmetic. The key identity:

(a + b) % m == ((a % m) + (b % m)) % m

This lets you avoid integer overflow by taking mod at each step.

Where it shows up: Pow(x, n) (LC 50), anything involving large numbers or "mod 10^9 + 7".

2. GCD and LCM

PYTHON

Copy

def gcd(a, b): while b: a, b = b, a % b return a lcm = lambda a, b: a * b // gcd(a, b)

Where it shows up: Fraction to Recurring Decimal (LC 166), water pouring problems.

3. Sieve of Eratosthenes

Find all primes up to n in O(n log log n). You won't implement this often, but recognizing when it applies saves time.

4. Fast Exponentiation

Compute x^n in O(log n) by repeated squaring:

PYTHON

Copy

def power(x, n): if n == 0: return 1 half = power(x, n // 2) return half * half * x if n % 2 else half * half

5. Geometry: Distance and Overlap

Most geometry problems reduce to:

• Distance: Compare squared distances to avoid floating point -- (x2-x1)^2 + (y2-y1)^2
• Rectangle overlap: Check if one rectangle is entirely left, right, above, or below the other
• Point in polygon: Ray casting (rarely asked, good to know exists)

When Interviewers Ask This

Classic Problems, Briefly

Pow(x, n) (LC 50) -- Fast exponentiation. Handle negative exponents by computing 1/x^|n|. Watch for integer overflow on n = INT_MIN.

Fraction to Recurring Decimal (LC 166) -- Long division simulation. Hash map detects when a remainder repeats -- that's where the repeating part starts.

Random Pick with Weight (LC 528) -- Build a prefix sum of weights, then binary search for a random target. Prefix sum converts weights into ranges on a number line.

K Closest Points to Origin (LC 973) -- Compare squared distances to avoid floating point. Use a max-heap of size k or quickselect.

Pattern Recognition Cheat Sheet

• "mod 10^9 + 7" --> modular arithmetic at every step
• "probability" or "random" --> prefix sums and binary search
• "coordinates" or "distances" --> geometry tricks, skip sqrt
• "how many" of something --> combinatorics (often DP)
• Tiny constraints (n at most 30) --> might be a math formula

Interview Tips

• Most math problems are medium difficulty. They rarely appear as the hard problem.
• If you don't see the trick, brute force simulation works for small inputs. Start there and optimize.
• Interviewers care about edge cases: negative numbers, zero, overflow, floating point precision.
• "Mod 10^9 + 7" is a signal that the answer is huge -- use modular arithmetic at every step, not just at the end.

Sources

• NeetCode Math & Geometry Roadmap -- curated problem ordering
• LeetCode Math Tag -- full problem list
• LeetCode Geometry Tag -- geometry-specific problems
• K Closest Points variants in Google interviews -- real interview discussion thread