Blog
Honest comparisons, interview tips and product news from the Ghost team.
SQL vs NoSQL: how to choose a database in a system design interview
"Why did you pick that database?" Saying "NoSQL because it scales" loses points. Here's how to justify the choice from the access patterns.
Vector database interview questions: embeddings, ANN indexes, filtering and scale
"Why is our vector search fast but missing obvious results?" Vector databases are now a standard AI engineer interview topic. Here are the questions and how to answer them.
Chatbot memory: how to make an LLM remember long conversations without blowing up cost
Forty messages in, your chatbot forgets the user's name. Here's why it happens, why a bigger context window only raises the bill, and how to answer the chatbot memory question in an AI engineer interview.
URL shortener system design: a step-by-step interview answer
"Design a URL shortener." It's the most common first system design question. Here's a clear answer, from requirements to scaling.
OFFSET vs cursor pagination: why page 10,000 is slow and how to fix it
Page one loads instantly. Page ten thousand times out. Same query. Here's why OFFSET gets slower the deeper you go, and the 5-step answer to the pagination interview question.
Consistent hashing explained: the hash ring, virtual nodes and how to answer it in an interview
"You have 4 cache servers and need to add a fifth without wiping the cache. How do you spread the keys?" Here's the answer, step by step.
AI interview copilot pricing compared (2026): what Cluely, Final Round AI, LockedIn, ParakeetAI and Ghost actually cost
Most interview copilots charge $66 to $299 a month. Here's what each one costs, what you get on the free plan, and what a few real interviews cost on each.
Cluely pricing (2026): every plan explained, and what it costs for interviews
Cluely's Pro plan is $19.99 a month, but staying hidden while you share your screen costs $149.99 a month. Here's what each plan includes and what it means for interviews.
Resume bullet points: a simple formula and 30 examples (with a free generator)
"Responsible for the checkout API" tells a recruiter nothing. Here's the formula for bullet points that show impact, with 30 before-and-after examples.
The hot key problem: when one Redis key takes down your cluster
A viral post sends every read to one cache key, and one Redis node melts while the rest sit idle. Here's why more nodes won't fix it, and how to answer the hot key interview question step by step.
Semantic caching for LLM apps: how it works and when it's safe (AI engineer interview)
"Half our support questions are the same 50 questions in different words. Can we stop paying the LLM to answer them again?" Here's how to answer it.
Database indexing interview questions: B-trees, composite indexes and query plans
"This query takes 4 seconds on a 50-million-row table. How do you fix it?" Here's how to answer, and the indexing questions that usually follow.
How to reduce LLM latency and cost: streaming, prompt caching and model routing
"Our AI feature takes 8 seconds and the bill doubled last month. What do you do?" A step-by-step answer for AI engineer interviews.
The Microsoft 365 outage: retry storms, cascading failures and how to answer the interview question
Outlook and Teams were down for over 12 hours. Here's what Microsoft has said so far, why servers that come back can fall over again, and how to answer "how do you stop a service from collapsing under load?"
The saga pattern: distributed transactions explained for backend interviews
"An order touches payments, inventory and shipping, each with its own database. What if inventory fails after the card is charged?" Here's the answer.
LLM tool calling: how to design an agent's tools (AI engineer interview)
"Design the tools for a customer support agent that can issue refunds." Here's how to answer it like someone who has shipped an agent.
Kafka vs RabbitMQ vs SQS: how to choose a message queue in an interview
"Why did you pick Kafka here and not RabbitMQ?" Most candidates say "Kafka is faster." Here's the answer that shows you understand the difference.
Caching strategies for system design interviews: cache-aside, write-through and invalidation
"Our product pages are slow and the database is at 90% CPU. Add a cache." Here's how to answer it, including the invalidation follow-ups interviewers always ask.
Flash sale system design: how to never sell stock you don't have
Two people bought the last phone and both saw "order confirmed". Here's the bug behind it, and the five-step answer interviewers want for flash sale system design.
Rate limiter system design: token bucket, sliding window and Redis, for interviews
"Design a rate limiter for a public API that runs on many servers." A classic system design question, with a five-step answer that covers the trade-offs interviewers probe.
Zero-downtime database migrations: the expand and contract answer for interviews
"Change a column on a table with hundreds of millions of rows, with zero downtime." Here's the expand-and-contract answer, and the details interviewers dig into.
LLM evals: how to catch regressions before users do (AI engineer interview)
Your team changes prompts and models every week. How do you know nothing got worse? Here's the eval setup interviewers expect you to describe.
How to debug a RAG system that gives wrong answers (AI engineer interview)
"Our RAG assistant gives confident wrong answers. How would you debug it?" The strong answer starts with one simple test that tells you where to look.
Idempotency keys explained: how to stop double charges, for backend interviews
The request timed out. Did the payment go through? Nobody knows, so the client retries. Here's how idempotency keys make that retry safe, and how to explain it.
How to scale a database: the backend interview answer, step by step
"Our database is the bottleneck. How would you scale it?" Most candidates jump straight to sharding. Here's the order interviewers actually want to hear.
Prompt injection in AI agents: how to answer it in an AI engineer interview
Your agent reads emails and web pages and can send messages. An interviewer asks how you'd stop prompt injection. Here's a layered answer that holds up to follow-up questions.
AI engineer resume guide: keywords, projects and bullet points that get interviews
Most AI engineer resumes list the same tools and the same tutorial projects. Here's how to write one that shows you can actually ship and measure LLM features.
STAR method examples for engineers: 8 behavioral questions with sample answers
Behavioral rounds reject more engineers than coding rounds. The STAR method, 8 questions you'll almost certainly get, and a sample answer for each.
"Tell me about yourself" for software and AI engineers: a formula and 4 examples
The first question of almost every interview, and the one most engineers wing. A simple formula and four sample answers you can adapt in 10 minutes.
LLM system design interview: a 7-step framework with a worked example
'Design a chatbot over our docs' sounds easy until the follow-ups start. A 7-step framework and a full worked example for LLM system design rounds.
25 RAG interview questions with answers, from basics to production
Retrieval-augmented generation is the most-asked topic in GenAI interviews. Here are the 25 questions that come up most, with answers that show real experience.
How to prepare for an AI engineer interview: a 4-week plan
Four weeks, one project, and a clear order to study in. A realistic plan for landing an AI engineer role, whether you're a fresher or switching from backend.
40 AI engineer interview questions (and how to answer them) for 2026
The 40 questions that keep coming up in AI engineer interviews, from tokens and attention to RAG, evals and cost, with short answers you can build on.
How to get your resume past ATS in 2026: a checklist that actually works
The ATS isn't secretly rejecting 75% of resumes. Here's what actually filters you out, and a 20-minute checklist to fix it.
AI interviewers are here: how to prepare when an AI interviews you
Your next interviewer might be an AI. The 3 types of AI interviews, what they actually score, and 10 practical ways to prepare.
Cluely charges $149.99 a month to hide from screen share. Ghost does it on the free plan.
Six AI interview copilots, their real prices and the fine print they hope you skip. One of them costs about $2 an interview. The rest want a subscription.