A software developer with a background in telecommunications and gamedev shares their one-week crash course preparing for a Google technical interview. With no prior algorithmic training, they used an LLM (Gemini Pro) as a personalized tutor to learn patterns across ~34 LeetCode problems in days. The post details their daily study protocol, key insights about algorithm patterns (graphs, trees, BFS/DFS, backtracking, two pointers), and the interview itself — where they successfully identified a graph+binary search hybrid problem but failed under stress to recall iterative binary search syntax. Despite coding errors, they were invited to on-site interviews. The post reflects on LLMs as tutors, the gap between algorithmic fluency and real-world coding, and whether speedrun learning is a valid preparation strategy.
Table of contents
IntroductionDay 0: Making the plan and procrastination dressed as learning.Day 1: Initial setup and warm-upDay 2: As much concept as possibleDay 3: Time to use learned knowledgeDays 4, 5, 6: Consolidating Acquired Knowledge Under a Hard DeadlineThe InterviewThe Cooling, Waiting TimeEpilogueSort: