Design ChatGPT - System Design Mock Interview (with eBay EM)

Make sure you're interview-ready with Exponent's system design interview prep course: https://bit.ly/3NxjDyT

Join Neeraj, an eBay engineering manager, as he builds ChatGPT during a system design mock interview. He identifies the requirements and constraints for building a robust chatbot system that can handle millions of users and transactions.

An overview of the high-level design of this system:

- Functional requirements:
- Create, update, view, and delete conversations
- Rate responses with thumbs up or thumbs down to train the model
- Text-based inputs in English
- Sanitization phase to remove profanity and detect insults

- Non-functional requirements:
- Latency, security, and scalability
- Login flows and rate limitation procedures to prevent DDoS attacks
- Scalable rate limiter and database storage

- High-level design:
- Conversation service manages conversation between users and chat GPT model
- Sanitization and analysis of messages before being sent to chat GPT model
- Results from chat GPT model are sent back for storage in database
- Thumbs down ratings
- Rest API with operations for creating, deleting, viewing, and sending messages
- Each conversation has a unique ID for each message
- NoSQL database storage where the conversation table contains various conversation IDs, and each message contains an ID, text, author, and parent

- Chat GPT model:
- Uses Transformer model to predict natural sequences of words
- Trained on internet data such as websites, books, and Wikipedia links
- Can use multiple approaches to select the most accurate prediction

- Chatbot design:
- Designed using supervised fine-tuning, a reward model, and reinforcement learning
- Reward model selects best responses to train the chatbot
- Chatbot can support different input and output formats, including images, audio, and video
- Takes into account the emotion and tone of the question and answer and will continuously train itself to improve accuracy

Watch more system design videos here:
Meta engineering manager answers a rate limiter interview question: https://youtu.be/SgWb6tWx3S8
Google SWE answers an algorithms interview question: https://youtu.be/NRRyk0XqkkA
Google TPM answers Tiktok system design interview question: https://youtu.be/Z-0g_aJL5Fw
Flipkart EM “Design Amazon Prime Video” system design interview question: https://youtu.be/PuU_0esYyhg

👉 Subscribe to our channel: http://bit.ly/exponentyt
🕊️ Follow us on Twitter: http://bit.ly/exptweet
💙 Like us on Facebook for special discounts: http://bit.ly/exponentfb
📷 Check us out on Instagram: http://bit.ly/exponentig
📹 Watch us on TikTok: https://bit.ly/exponenttikttok

ABOUT US:
Want to land your dream career in tech? Exponent is an online community, course, and coaching platform to help you ace your upcoming system design interviews.

Exponent has helped hundreds of thousands of candidates pursue and land their dream careers at companies like Google, Microsoft, Amazon, and high-growth startups. Exponent is currently licensed by Stanford, Yale, UW, and others.

Our system design courses include interview lessons, questions, and complete answers with video walkthroughs. Get access to hours of real interview videos, where we analyze what went right or wrong, and our community of expert coaches and industry professionals to help you get your dream job and more!

Видео Design ChatGPT - System Design Mock Interview (with eBay EM) канала Exponent