Skip to contentSkip to editor
Docs

Getting Started with Sophocode

Learn how to use Sophocode to practice coding interviews with Sophia, your wise AI coach. No sign-up required.

Browse docs

What is Sophocode?

Sophocode is an AI-powered coding interview practice platform built around one idea: wisdom over grinding.

The name comes from the Greek sophos (σοφός) — wisdom. Where LeetCode-style grinding builds muscle memory, Sophocode builds understanding. You'll practice with Sophia, your AI coach who guides you through the Clarify → Plan → Code → Reflect process.

How a Session Works

Every practice session follows the Clarify → Plan → Code → Reflect loop:

  1. Clarify — Read the problem and ask clarifying questions. Sophia will prompt you if you skip this step.
  2. Plan — Describe your approach before coding. Identify the pattern, state your complexity, and outline the algorithm.
  3. Code — Write your solution in the browser editor. Python runs in your browser via Pyodide — no server needed.
  4. Reflect — Review your solution. Sophia gives feedback on your approach, complexity, and what to work on next.

Choosing a Problem

From the Practice page, you can browse problems filtered by:

  • Pattern (Sliding Window, Two Pointers, etc.)
  • Difficulty (Easy, Medium, Hard)

If you're new, start with an Easy problem in a pattern you're curious about. The goal isn't to solve the hardest problems — it's to build pattern recognition from the ground up.

Getting Hints

Every problem has 3 progressive hints:

  • Hint 1 — Points you toward the right pattern
  • Hint 2 — Gives you the key insight or data structure
  • Hint 3 — Provides the algorithm sketch

Wait at least 10-15 minutes before using a hint. The struggle before the hint is where learning happens.

Your Progress

Sophocode tracks:

  • Which problems you've attempted and solved
  • How many hints you needed
  • Your mastery level per pattern
  • Problems scheduled for spaced-repetition review

You don't need an account to start practicing. Anonymous session identity is managed with a secure cookie and you can create an account later to link your progress.

Pattern Taxonomy

PatternWhat It Solves
Two PointersPair finding, partitioning in sorted arrays
Sliding WindowLongest/shortest subarray/substring conditions
Fast & Slow PointersCycle detection, middle of linked list
Merge IntervalsOverlapping intervals, scheduling
Cyclic SortMissing/duplicate numbers in [1,n] range
In-Place ReversalReversing linked lists without extra space
Tree BFSLevel-order traversal, shortest paths
Tree DFSPath problems, subtree problems
Two HeapsMedian tracking, balanced partitions
Subsets/BacktrackingAll combinations, permutations, subsets
Modified Binary SearchRotated sorted arrays, searching on answer
Top K ElementsK largest/smallest/most frequent
K-way MergeMerging K sorted lists/arrays
Dynamic ProgrammingOptimization, counting with overlapping subproblems