Protocols

Before The Interview

• As soon as the interview is scheduled, request as much information as possible about the interview from the recruiter. They may not be able to tell you everything, but every bit counts!

  Who will be interviewing me? What is their name or what team/product do they work on?
  
  How many problems do I have to solve?
  
  What will the difficulty of the problems be?
  
  Are there any particular topics I should expect?
  
  Should I prioritize solving all problems in time, or explaining my thoughts and writing clean code?
  

• Do a mock interview of the same format your interview will be: Behavioral, Algorithms, or System Design

• Practice algorithmic problems while simulating real interview conditions.

  • Give yourself 20m per question
  • Stick to 5m time blocks for each phase of the problem: Analysis → Design → Implementation → Complexity Analysis and Optimizations
  • If your interview will be on a whiteboard, solve the problems on paper.
  • Try to explain your thought process out loud or in comments, to simulate a discussion with your interviewer.

• After a practice problem or mock technical interview, use this rubric to evaluate your performance.

  Points	Task
  3	Asking clarifying questions, discussing edge cases, and planning your solution.
  3	Communicating what you're doing and having a back-and-forth dialogue with the interviewer.
  3	Getting a working solution.
  3	Time/space complexity analysis and discussing possible optimizations (if any).
  3	Overall code cleanliness and adhering to coding conventions.

During The Interview

1. [5m] Analyze the problem. Even if you don’t know how to solve the problem, focus on showing the interviewer everything you DO know.

   • Ask clarifying questions to understand problem.

   • Ask about edge cases and assumptions being made.

     • Character set size (e.g. ASCII = 128, 256, Unicode = 221)
     • Binary number size (32-bit)
     • Input dataset size range.

   • Mention any patterns you recognize from the problem keywords or requirements.

     Keyword / Requirement	Pattern
     Filter out Duplicates	HashSet
     Tracking frequency of String characters / Array values	HashMap
     Kth Largest/Smallest/Most frequent	Heap, QuickSelect
     List out all Possibilities	Backtracking
     Sorting an Array with Bounded values	BucketSort
     Given a Sorted Array	Binary Search, Two Pointers
     Storing and Searching for Timestamps	Binary Search
     Cycles, Components, or Redundant Edges in a Graph	Union Find
     Log(N) Time Requirement	Binary Search
     Working with Parentheses / Order of Operations	Stack
     Checking contiguous Substrings/Subsequences	Sliding Window
     Checking noncontiguous Subsequences	DP, HashSet
     Searching for Word Prefixes	Trie
     O(1) Space Requirement + Math/Number Problem	Bit Manipulation
     Input can be represented as a Directed Graph	Topological Graph
     Shortest Path in a Graph	BFS, Multisource BFS

   • Explain the best conceivable runtime for the problem, and why that is

     • Unsorted input, sorted result - $O(NLogN)$
     • If you have to look at every element in the dataset at least once - $O(N)$

   • Brainstorm potential solutions

     • Start small. Discuss the Brute Force solution. Explain what it would be, it’s Time and Space complexity, and why it’s inefficient.
     • Mention relevant data structures and algorithms that might be useful for this problem, why they would be useful, and the complexity of key operations
       • HashMaps have O(1) / constant search
     • If you can think of the optimal solution out of the box, explain that next. If not, talk about potential optimizations to the brute force solution. Explain everything you’re considering.
       • Look for repeated work (tell them you’re doing this).
       • Go down the Quick Optimization Checklist below.
2. [5m] Design your solution.
   • Choose the best solution you could come up with.
   • Write pseudocode for it.
   • Walk through an example with your pseudocode.
3. [5m] Implement your solution.
   • Write clean code.
     • Use short, self-explanatory names.
     • Write subproblems / independent steps as one-line function calls in the main function first. Then implement the functions later.
   • Add edge cases:
     • Null input
     • Size N = 0, 1 input
     • Negative input
4. [5m] Complexity Analysis and Optimizations
   • Explain the time and space complexity of your solution.
   • Discuss potential further optimizations (don’t need to implement).
   • Make this a back and forth with your interviewer.

Coding Tips

Choosing a Language

• Pick whichever language your most comfortable in.
• If possible, a concise language like Python is preferable to a more verbose one like Java. This way you’ll spend less time writing out boilerplate code during the interview.

Writing Code on a Whiteboard

• Start at the top-left of whiteboard, and avoid writing slanted lines. This will take some practice.
• Leave ample extra space in between lines, so that you can go back later in case you missed something.