Ace The ML System Design Interview [Part 5] - Modeling

Modeling plays a critical role in machine learning system design, with classification and regression being the most frequently encountered problem types in interviews. Here are the core points from this lesson:

- Understand various evaluation metrics, such as precision, recall, F1 score, NDCG, and loss metrics like cross-entropy and focal loss, and know when to apply each
- Be prepared to discuss trade-offs between metrics, particularly precision and recall, and avoid over-optimizing for a single metric
- Recognize the significance of loss metrics, memorizing the equations for cross-entropy and focal loss, and understanding their applications in balanced and imbalanced datasets
- Be ready to discuss model selection, including architecture choices like two-tower models or matrix factorization, and anticipate follow-up questions on optimization techniques and troubleshooting issues like overfitting and vanishing gradients
- Consider a baseline model for comparison if no current system exists, and if there is an existing system, use it as a reference point

If you like what Ilya has to say, subscribe to his YouTube for more high-quality ML/AI career guidance: [MLEpath - YouTube](https://go.mlepath.com/taro-youtube-mlepath)

If you want hands-on support from Ilya to crack the FAANG ML interview, join his coaching program: [MLEpath - Coaching Program](https://go.mlepath.com/taro-march-25)

Ace The ML System Design Interview [Part 4] - Data

Data is the foundation of machine learning, and a thorough discussion of data is essential in any machine learning system design interview. Here are the core points from this lesson:

- Discuss labels, features, and data set splitting, understanding the trade-offs involved in each area
- Explore various methods for generating data labels, such as human annotation, synthetic data, and LLMs, while considering the pros and cons of each approach
- Focus on high-level feature sources, providing relevant examples and selecting predictive features while avoiding sensitive topics
- Understand feature encoding techniques, such as one-hot encoding and embeddings, and their impact on model performance
- Weigh the trade-offs between cross-validation and train-validate-test splits, while considering issues like data leakage and imbalances in the data set

If you like what Ilya has to say, subscribe to his YouTube for more high-quality ML/AI career guidance: [MLEpath - YouTube](https://go.mlepath.com/taro-youtube-mlepath)

If you want hands-on support from Ilya to crack the FAANG ML interview, join his coaching program: [MLEpath - Coaching Program](https://go.mlepath.com/taro-march-25)

Ace The ML System Design Interview [Part 1] - Introduction

We've created this course to help you navigate the machine learning system design interview, drawing from our extensive experience in big tech. Our goal is to demystify the process and provide you with the tools to succeed. Here's what you'll learn from this course:

- A structured approach to tackling complex machine learning system design interviews
- How to showcase your machine learning expertise, seniority, and communication skills beyond just designing a system
- Techniques for effectively communicating complex machine learning concepts to both technical and non-technical stakeholders
- Strategies to evaluate trade-offs and analyze the pros and cons of different engineering techniques
- How to stay calm, focused, and confident throughout the interview process

If you like what Ilya has to say, subscribe to his YouTube for more high-quality ML/AI career guidance: [MLEpath - YouTube](https://go.mlepath.com/taro-youtube-mlepath)

If you want hands-on support from Ilya to crack the FAANG ML interview, join his coaching program: [MLEpath - Coaching Program](https://go.mlepath.com/taro-march-25)

Ace The ML System Design Interview [Part 3] - Big Picture Design

A strong big-picture system design is crucial for success in machine learning system design interviews. It's not about drawing a perfect diagram immediately, but about demonstrating an understanding of the system's components and their interactions. Here are the core points from this lesson:

- Identify the inputs and outputs of the system during the alignment phase to establish a foundation for building the diagram
- Start with the user and the service, then progressively add components, especially those involving machine learning, to illustrate the flow of data and processes
- Break down machine learning components beyond a single "ML system" box to show the different stages and types of models involved
- Include essential infrastructure elements, such as feature stores, to demonstrate a comprehensive understanding of the system
- Iterate on the diagram, adding or modifying components as a deeper understanding of the problem and its requirements develops

If you like what Ilya has to say, subscribe to his YouTube for more high-quality ML/AI career guidance: [MLEpath - YouTube](https://go.mlepath.com/taro-youtube-mlepath)

If you want hands-on support from Ilya to crack the FAANG ML interview, join his coaching program: [MLEpath - Coaching Program](https://go.mlepath.com/taro-march-25)

Ace The ML System Design Interview [Part 2] - Why You Need A System

The machine learning system design interview can be daunting, but understanding common pitfalls can help you succeed. Here are the core points from this lesson:

- Align with the interviewer on the problem statement before diving into design to ensure the right problem is being solved
- Use a structured approach: start with big-picture system design, then address data considerations and modeling, only discussing deployment and operations if time allows
- Clearly articulate reasoning by explaining the "why" behind design choices and considering trade-offs between different techniques
- Take control of the interview by making assumptions, validating them with the interviewer, and simplifying the problem to fit within the given time
- Write down key assumptions and decisions to maintain clarity and provide a reference for both yourself and the interviewer

If you like what Ilya has to say, subscribe to his YouTube for more high-quality ML/AI career guidance: [MLEpath - YouTube](https://go.mlepath.com/taro-youtube-mlepath)

If you want hands-on support from Ilya to crack the FAANG ML interview, join his coaching program: [MLEpath - Coaching Program](https://go.mlepath.com/taro-march-25)

Taro Book Club: The Pragmatic Programmer - Session 6

Session Led by [Bikram Kashyap](https://www.linkedin.com/in/kashyapbikram/)

In this #taro-bookclub discussion, we'll go over [**"The Pragmatic Programmer"**](https://www.amazon.com/Pragmatic-Programmer-Journeyman-Master/dp/020161622X) by Andrew Hund & David Thomas

Here's the agenda:

- Design by Contract
- Dead Programs Tell No Lies
- Assertive Programming
- How to Balance Resources
- Don't Outrun Your Headlights

Please send your questions ahead of time to [**Bikram Kashyap**](https://www.linkedin.com/in/kashyapbikram/) in the [**Taro Premium**](https://www.jointaro.com/membership/) Slack, so that they can be included in the discussion. *If you don't have the book or want more details on how this all works, check out the #taro-bookclub channel in the Taro Premium Slack as well!*

Get ready for another engaging and informative session!

Part 8 [RBP] Next Steps With React

Follow Florian on various social media platforms like [YouTube](https://www.youtube.com/c/CodinginFlow), [Instagram](https://www.instagram.com/codinginflow/), [TikTok](https://www.tiktok.com/@codinginflow), and [X/Twitter](https://x.com/codinginflow) for more web development content and updates. Mailing list: [CodingInFlow.com/newsletter](http://CodingInFlow.com/newsletter)

Links:

- Florian’s free web development playlist on YouTube: <https://www.youtube.com/playlist?list=PLrnPJCHvNZuDG6BWjUFVxeWyYlve_SVJO>
- Code link: <https://github.com/codinginflow/react-best-practices/branches/stale>

Part 7 [RBP] Use Fragments Instead Of Divs

This lesson focuses on using React Fragments to improve the structure and styling of your components when rendering multiple elements.

- **The Problem with Divs:** Learn how wrapping elements in unnecessary divs for rendering can lead to cluttered DOM structures and styling issues.
- **Introducing Fragments:** Discover how to use Fragments (empty tags `<> </>`) to group elements without adding extra nodes to the DOM, resulting in cleaner and more efficient code.
- **Benefits of Fragments:** Understand how Fragments help maintain proper styling, improve code readability, and optimize your application's performance by reducing unnecessary DOM elements.

Links:

- Florian’s free web development playlist on YouTube: <https://www.youtube.com/playlist?list=PLrnPJCHvNZuDG6BWjUFVxeWyYlve_SVJO>
- Code link: <https://github.com/codinginflow/react-best-practices/blob/fragments/src/app/page.tsx>

Part 6 [RBP] Create Hooks For Your Context Providers

This lesson presents a best practice for working with context providers in React, specifically addressing situations where a meaningful default value isn't available.

- **The Problem with Undefined Context:** Understand the challenges of using context providers without a proper default value, leading to potential silent failures and difficulties in debugging.
- **Custom Hooks for Context Providers:** Learn how to create custom hooks that encapsulate the logic for accessing and validating context values, ensuring that errors are caught early in the development process.
- **Improving Code Reliability and Reusability:** Discover how this approach improves code readability, promotes reusability of context-related logic, and provides more informative error messages, making your React applications more robust.

Links:

- Florian’s free web development playlist on YouTube: <https://www.youtube.com/playlist?list=PLrnPJCHvNZuDG6BWjUFVxeWyYlve_SVJO>
- Code link: <https://github.com/codinginflow/react-best-practices/tree/context-hook/src/app>

Part 5 [RBP] Create Custom Hooks

This lesson explains how to improve the organization and reusability of React components by extracting complex logic, particularly related to data fetching, into custom hooks.

- **Creating Custom Hooks:** Learn the process of creating a custom hook to encapsulate state management and side effects, such as those involved in fetching data and handling loading and error states.
- **Organizing and Reusing Logic:** Understand the benefits of separating concerns by moving complex logic into custom hooks, leading to cleaner and more maintainable components.
- **Improving Code Readability:** Discover how custom hooks can make your code more declarative and easier to understand by abstracting away implementation details and providing a clear interface for interacting with data.

Links:

- Florian’s free web development playlist on YouTube: <https://www.youtube.com/playlist?list=PLrnPJCHvNZuDG6BWjUFVxeWyYlve_SVJO>
- Code link: <https://github.com/codinginflow/react-best-practices/tree/custom-hook/src>

Part 4 [RBP] Clean Up Your Fetch Requests

This lesson focuses on preventing race conditions that can occur when making multiple fetch calls in React components, leading to inconsistent UI behavior and a poor user experience.

- **Identifying Race Conditions:** Understand how race conditions happen when asynchronous fetch requests resolve in an unexpected order, particularly with slow or unreliable network connections.
- **Strategies for Avoiding Race Conditions:** Learn how to use tools like SWR or React Query to manage fetch requests and automatically handle race conditions, or implement manual cleanup logic with an "ignore" flag to discard outdated responses.
- **Improving Code Structure:** Discover how to keep your fetch-related code organized and maintainable, especially when dealing with loading and error states, setting the stage for cleaner solutions covered in the next lesson.

Links:

- Florian’s free web development playlist on YouTube: <https://www.youtube.com/playlist?list=PLrnPJCHvNZuDG6BWjUFVxeWyYlve_SVJO>
- Code link: <https://github.com/codinginflow/react-best-practices/blob/fetch-cleanup/src/app/page.tsx>

Part 3 [RBP] Clean Up Your Effects

This lesson emphasizes the crucial practice of cleaning up side effects in React components to prevent unexpected behavior and optimize performance.

- **Understanding useEffect Cleanup:** Learn how to use the cleanup function returned by `useEffect` to properly handle resources like network connections, event listeners, and subscriptions.
- **Avoiding Memory Leaks and Unexpected Behavior:** Discover how neglecting cleanup can lead to memory leaks, multiple event triggers, and other unintended consequences, especially when components mount and unmount frequently.
- **Debugging with Strict Mode:** Explore how React's Strict Mode helps identify potential issues by intentionally double-invoking effects in development, making it easier to spot missing cleanup logic.

Links:

- Florian’s Next.js with Express and TypeScript Course: <https://codinginflow.com/nextjs>
- Florian’s free web development playlist on YouTube: <https://www.youtube.com/playlist?list=PLrnPJCHvNZuDG6BWjUFVxeWyYlve_SVJO>
- Code link: <https://github.com/codinginflow/react-best-practices/blob/effect-cleanup/src/app/page.tsx>

Part 2 [RBP] Derive State When Possible

This lesson focuses on optimizing React state management for better performance and code organization.

- **Deriving State:** Learn how to calculate and derive state values directly within your component instead of storing redundant data, leading to more efficient and maintainable code.

- **Avoiding Unnecessary Re-renders:** Understand the impact of state changes on component re-renders and discover techniques to minimize them, such as using `useMemo` for expensive calculations.

- **Improving Code Clarity:** Explore how to simplify your state logic and reduce complexity by deriving state and avoiding unnecessary `useEffect` calls.

- Florian’s free web development playlist on YouTube: <https://www.youtube.com/playlist?list=PLrnPJCHvNZuDG6BWjUFVxeWyYlve_SVJO>

- Code link: <https://github.com/codinginflow/react-best-practices/blob/derive-state/src/app/page.tsx>

Part 1 [RBP] Exploring Modern React Frameworks + Setup

This lesson covers setting up a new React project using best practices. It discusses choosing the right framework, emphasizing the benefits of Next.js for its full-stack capabilities and features like routing and image optimization. The lesson also touches on the importance of TypeScript and ESLint for code quality and maintainability, and introduces the App Router as the modern approach to structuring Next.js applications.

- Florian’s free web development playlist on YouTube: <https://www.youtube.com/playlist?list=PLrnPJCHvNZuDG6BWjUFVxeWyYlve_SVJO>
- Code link: <https://github.com/codinginflow/react-best-practices/branches/stale>

Part 0 [RBP] React Best Practices - Course Intro

Florian Walther, from Coding In Flow, shares the common pitfalls and best practices in React development. This course will help developers write cleaner and more robust React code by addressing frequent mistakes. The lesson also highlights additional resources available, including a YouTube channel, newsletter, and a GitHub repository with code examples.

- Florian’s free web development playlist on YouTube: <https://www.youtube.com/playlist?list=PLrnPJCHvNZuDG6BWjUFVxeWyYlve_SVJO>
- Code link: <https://github.com/codinginflow/react-best-practices/branches/stale>

Paper Reading: DeepSeek R1 - Incentivizing Reasoning Capability in LLMs via Reinforcement Learning

Session led by [Mike A](https://www.linkedin.com/in/malik-mike-a-32b63818b/) 

Join us for an insightful session on the groundbreaking paper, DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning. In this session, we will explore DeepSeek-R1, a state-of-the-art reasoning model that pushes the boundaries of reinforcement learning (RL) applied to large language models (LLMs). 

This paper introduces two novel models, DeepSeek-R1-Zero and DeepSeek-R1, showcasing advancements in reasoning capabilities through RL-driven self-evolution. Unlike traditional models relying heavily on supervised fine-tuning (SFT), DeepSeek-R1-Zero develops its reasoning abilities purely through RL, while DeepSeek-R1 combines RL with a multi-stage training pipeline for enhanced performance. 

The paper highlights exceptional benchmarks achieved by DeepSeek-R1 on math, coding, and STEM-related reasoning tasks, where its performance rivals that of leading closed-source models such as OpenAI's o1-1217. Additionally, we’ll discuss distilling these capabilities into smaller, more efficient models to make advanced reasoning accessible for diverse applications. 

Whether you're a researcher, developer, or enthusiast in AI and LLMs, this paper reading will provide an in-depth understanding of the novel reinforcement learning techniques driving DeepSeek-R1 and its implications for the future of AI-driven reasoning systems. Don't miss this opportunity to engage with cutting-edge advancements in the field! \[2501.12948\] DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning 

GitHub - deepseek-ai/DeepSeek-R1 

Mark Chen X Post: <https://x.com/markchen90/status/1884303237186216272?s=12&t=BSlMYtugAr8LLWXLnMLLrQ> 

GRPO Diagram: <https://www.reddit.com/r/LocalLLaMA/comments/1i78sfs/deepseek_r1_grpo_code_open_sourced/#lightbox> 

The Bitter Lesson: <http://www.incompleteideas.net/IncIdeas/BitterLesson.html> 

Gradient Descent Image: <https://miro.medium.com/v2/format:webp/1\\\*f9a162GhpMbiTVTAua_lLQ.png> 

V3: <https://arxiv.org/abs/2412.19437> 

DeepSeek Math: <https://arxiv.org/abs/2402.03300> 

rstar-math: <https://arxiv.org/abs/2501.04519> 

Verify Step-by-Step: <https://arxiv.org/abs/2305.20050> 

DeepSeek catch-up chart: <https://media.licdn.com/dms/image/v2/D4E22AQG552O63UNJWg/feedshare-shrink_800/B4EZTN\\\[…\\\]41824000&v=beta&t=ubyKQniaCJTL37PzIOJi9YZRo1AF8yipuauSioyn59U> 

UC Berkeley Student $30 replication of Aha: <https://x.com/jiayi_pirate/status/1882839370505621655> 

Deep Agent R1-V replication: <https://x.com/liangchen5518/status/1886171667522842856> 

s1: <https://arxiv.org/abs/2501.19393> 

DeepSeek-r1 cost breakdown: <https://semianalysis.com/2025/01/31/deepseek-debates/> 

R1 Deep Dive: <https://fireworks.ai/blog/deepseek-r1-deepdive> 

All in Podcast: <https://podcasts.apple.com/us/podcast/all-in-with-chamath-jason-sacks-friedberg/id1502871393?i=1000687638652> 

Peter Gostev LinkedIn: <https://www.linkedin.com/in/peter-gostev/> 

LLM Agents Learning: <https://llmagents-learning.org/sp25>

Engineer Your Ideal Tech Career [Part 17] - Closing Thoughts

Wrapping up the course, [Sophie](https://www.linkedin.com/in/sophienovati/) shares her motivations behind helping people achieve their most ideal and ambitious career outcomes and some ways to connect further.

We hope you leave this course with a crystal clear vision of what brings you excitement as an engineer and how to take that next step to get there.

Thanks for watching!

Looking for mentorship from tech’s top engineers + personalized job hunt assistance and interview prep? [Formation.dev](http://Formation.dev) can provide you unlimited career coaching, technical mentorship, personalized practice, and mock interviews. Join us here: <https://formation.dev/partners/taro>

Worksheet: [Formation - Coding Your Career](https://firebasestorage.googleapis.com/v0/b/tech-career-growth.appspot.com/o/courses%2Fslides%2FFormation%20Career%20Direction%20Worksheet.pdf?alt=media&token=f6344b78-529a-411e-a449-7a34a8d55fc8)

Engineer Your Ideal Tech Career [Part 16] - Specialization

Now, depending on which specialization you’ve landed on, your near-term next steps and interview expectations also vary.

In your current role, the advice may be to start shipping easter eggs, optimizing for teams with grand scale, or rapidly learning new technologies.

Should you be thinking about a change soon, there are also actionable interview prep tips, particularly around system design, (and mistakes to avoid!) in this section.

Looking for mentorship from tech’s top engineers + personalized job hunt assistance and interview prep? [Formation.dev](http://Formation.dev) can provide you unlimited career coaching, technical mentorship, personalized practice, and mock interviews. Join us here: <https://formation.dev/partners/taro>

Worksheet: [Formation - Coding Your Career](https://firebasestorage.googleapis.com/v0/b/tech-career-growth.appspot.com/o/courses%2Fslides%2FFormation%20Career%20Direction%20Worksheet.pdf?alt=media&token=f6344b78-529a-411e-a449-7a34a8d55fc8)

Engineer Your Ideal Tech Career [Part 15] - Role Types

Depending on what role you’ve narrowed your sights on, there are a number of things you can do in your current role to begin charting progress — from picking a specialization, to start scoping projects and reviewing code, to volunteering for mentorship opportunities and hiring.

Should you be thinking about a change soon, there are also actionable interview prep tips (and mistakes to avoid!) in this section.

Looking for mentorship from tech’s top engineers + personalized job hunt assistance and interview prep? [Formation.dev](http://Formation.dev) can provide you unlimited career coaching, technical mentorship, personalized practice, and mock interviews. Join us here: <https://formation.dev/partners/taro>

Worksheet: [Formation - Coding Your Career](https://firebasestorage.googleapis.com/v0/b/tech-career-growth.appspot.com/o/courses%2Fslides%2FFormation%20Career%20Direction%20Worksheet.pdf?alt=media&token=f6344b78-529a-411e-a449-7a34a8d55fc8)

Engineer Your Ideal Tech Career [Part 14] - Defining Near-Term Goals

Moving into your last section and with inspiration from the engineers you just heard from, you can expect to start defining near-term goals that you can implement within the next week, month, or quarter, to start charging towards your ideal long-term career.

This section will include tangible advice you can implement on the job or if you are preparing for interviews.

Looking for mentorship from tech’s top engineers + personalized job hunt assistance and interview prep? [Formation.dev](http://Formation.dev) can provide you unlimited career coaching, technical mentorship, personalized practice, and mock interviews. Join us here: <https://formation.dev/partners/taro>

Worksheet: [Formation - Coding Your Career](https://firebasestorage.googleapis.com/v0/b/tech-career-growth.appspot.com/o/courses%2Fslides%2FFormation%20Career%20Direction%20Worksheet.pdf?alt=media&token=f6344b78-529a-411e-a449-7a34a8d55fc8)

Tech Career Growth Videos

Our instructional videos provide software engineers with insights needed for career growth. Masterclasses go incredibly in-depth on a specific tech topic while case studies are real examples from established engineering leaders.