Kaloyan Kraev

🚀 How to Approach a System Design Interview: Amazon Prime Video Example When tackling a system design problem, especially for a complex platform like Amazon Prime Video, it's crucial to approach it strategically. Here's how I break it down: Get clarity on the requirements, both functional and non-functional. Focus on the user flows, define the scale, detail out the components, improve scalability, and wrap up confidently. Pro Tip for Interviews: Don’t rush into drawing diagrams! Ask clarifying questions, understand the problem fully, and then build the system piece by piece. Communicate your thought process clearly, and focus on scalability, reliability, and cost-efficiency. Remember, the trick is to think aloud—explain why you're making certain decisions. The interviewer wants to understand your reasoning as much as the final solution! Good luck to all those preparing for system design interviews! You've got this!💪 Want to master system design interviews and land those $500k+ TPM roles? Make sure you're fully prepared with the Kracd.com interview prep course. Sign up today and transform your career in just 30 days. #SystemDesign #AmazonPrimeVideo #TechInterview #InterviewTips #Scalability #CDN #TechCareers #SoftwareEngineering #Microservices #TechLeadership

9

I am excited to share the latest blog, co-authored by myself, Junhwi Kim and Vyas Sekar, detailing how we managed to successfully localized and improved LLM by combining Fine-Tuning and RAG. https://lnkd.in/gKDkgGPq

70

3 Comments

An open letter from Meta employees to Mark Zuckerberg and Meta's leadership. Very well articulated, I couldn't have done it myself better! I left Meta end of 2023 because of many things similar to what's mentioned in the letter, so I can definitely relate.

88

The Lesson From the Movie "Dor" for Life Design There’s an unforgettable scene in Dor where Zeenat, against all odds, embarks on a journey to seek forgiveness from Meera. She has no guarantees, no assurances, and no clear path. But somehow, she finds her way to Meera’s heart. Somehow, she creates a bond of love and understanding that transforms their lives. I found out the code to life design Zeenat used that helps her navigate uncertainty and create miracles. And this Code is Just 3 Letters: "How" See, the reason most people struggle to design the life they desire is that they’re stuck in "Why" questions: Why is life so unfair to me? Why does it seem impossible to achieve my dreams? But the code to life design, as I learned from Zeenat, is this: Moving from "Why" to "How" faster than anyone else. Instead of asking "Why?" she focuses on "How?" How can I find the courage to face my challenges? How can I create a connection where none seems possible? No matter the obstacle, Zeenat never stays stuck in "Why." She takes the first step, trusting that somehow the answers will emerge along the way. That’s why we love her character. That’s how we design extraordinary lives as well. So, will you move from "Why" to "How" fast, like Zeenat does in Dor? Type "How" in the comments if you understand the Dor Code to life design. #sanghamitaroychoudhury #shebundant #lifebydesignnotbydefault #choices #How #challengechallengetoday #success #freedomlifestyle

1

From a Garage to a Global Giant: Lessons from Amazon’s Early Days A small garage in Bellevue, Washington. That’s where Jeff Bezos and a handful of employees worked tirelessly on a website called Cadabra (yes, that was Amazon’s original name). Their vision? To create an online bookstore—a novel concept in a world where buying books meant visiting a physical store. But here’s the thing: their journey was far from glamorous. Lesson 1: Dream Big, Start Small Amazon didn’t start by selling everything under the sun. Instead, Bezos focused on books because they were easy to ship and had an enormous catalog. The goal wasn’t just to sell books—it was to master e-commerce logistics and customer experience. As Bezos famously said, "We’re not a bookstore. We’re a tech company that just happens to sell books." Lesson 2: Embrace Challenges In its early days, Amazon faced hurdles that would have crushed many startups. From convincing investors about online retail (a risky bet in the '90s) to operating on razor-thin margins, Bezos and his team had every reason to quit. But instead of running from challenges, they leaned into them. One story stands out: Amazon's employees would personally pack and ship books. It wasn’t just about saving costs—it was about building a company culture that prioritized getting things done over fancy titles or comfort. Lesson 3: Obsess Over Customers Bezos instilled a customer-first philosophy from day one. When Amazon launched its reviews feature, publishers protested: “Why would you allow negative reviews?” Bezos stood firm. He believed that trusting customers would eventually build loyalty—and he was right. This obsession with customers is why Amazon grew from a simple online bookstore to the everything store we know today. Why This Matters Amazon’s story isn’t just about innovation or strategy—it’s about resilience, adaptability, and an unwavering belief in the power of the internet to transform commerce. Whether you’re building your first project, leading a team, or managing a product, Amazon’s journey reminds us that big dreams often have humble beginnings. So here’s my takeaway for you: 👉 Start where you are, with what you have. 👉 Solve one problem at a time, but keep your eyes on the long-term vision. 👉 And never stop obsessing over the people you serve—your users, your customers, your team. #Inspiration #Entrepreneurship #Leadership #Innovation #AmazonStory

3

Hello, colleagues, this Yahoo article is interesting and relatable for professionals. Here is a teaser: "Amazon has an incredibly granular list of criteria that need to be met before you can be considered for promotion, so I worked with my manager to identify the gaps in my performance, ... and I created a project for each gap that, if completed, could serve as evidence of my ability. I worked backward to create a strict timeline to finish my projects ... that helped me get systematic about my approach moving forward." https://lnkd.in/dsfvPeAk Take a lesson from this person. I seems Amazon gets very analytical when it comes to promotions. And in a company that size, it is impossible to not promote without rigid check-off type policies or whatever XYZ job promotion software. But, hey, there is a lot I will do for a $500,000 salary. Please take note of this last sentence, if anyone works for Amazon and wants a Quality Engineering Manager/Director anywhere in the world.

1

1 Comment

𝗠𝗶𝗹𝗹𝗲𝗻𝗻𝗶𝗮𝗹𝘀 𝗶𝗺𝗺𝗶𝗴𝗿𝗮𝗻𝘁𝘀 𝗳𝗿𝗼𝗺 𝗜𝗻𝗱𝗶𝗮 — 𝗬𝗼𝘂 𝗵𝗮𝘃𝗲 𝗮𝗻 𝗮𝗱𝘃𝗮𝗻𝘁𝗮𝗴𝗲 𝗶𝗻 𝘁𝗵𝗲 𝗴𝗿𝗲𝗲𝗻 𝗰𝗮𝗿𝗱 𝗽𝗿𝗼𝗰𝗲𝘀𝘀. 𝗔𝗿𝗲 𝘆𝗼𝘂 𝗺𝗮𝗸𝗶𝗻𝗴 𝘁𝗵𝗲 𝗺𝗼𝘀𝘁 𝗼𝗳 𝗶𝘁? Meet Pooja, a Product Lead who moved to the U.S. in 2008 for her master’s. After graduation, she landed her first job in Boston, but her initial employer didn’t file for her green card. She later switched to a Fortune 500 company, which filed her EB2 in 2012. Now, over a decade later, Pooja is still navigating the H1-B renewal cycle. Meanwhile, junior colleagues who arrived years after her are securing green cards through EB-1. Feeling stuck, she’s considering EB-1 herself but isn’t sure if the process is worth it. What Pooja doesn’t realize: ➡️ Her profile may already be EB1-worthy. ➡️ She likely meets the Critical Role and Salary criteria. ➡️ Her contributions, speaking engagements, and professional growth over the years all count. With a few targeted efforts, she could have an EB1-ready profile within a year, port her 2012 priority date, and get her green card in 12-18 months. Now, Pooja has two choices: Option 1: Take control with EB-1. Option 2: Hope the EB2 backlog advances and gamble on the wait. If you were Pooja, what would you choose? PS: If Option 1 feels like your path, consider joining our EB1 Slack Community(link in comments)! Every month, we help at-least five people accelerate their EB1 journey with the right strategies and tools.

4

1 Comment

Re-posting one of my less popular / unorthodox views. Weekly 1:1s are hacks to get around the lack of proper team communication setups, and they weaken individual resilience. Adhoc 1:1s are fine for providing spot feedback, but they should be rare. Career growth 1:1s are great, but they should happen once in 3-6 months. Project check-ins / questions should be discussed in 5-10 minutes on slack. A design whiteboarding meeting between two engineers is fantastic, but it's not a 1:1. Weekly 30 minute 1:1s are unhealthy and an unnecessary time-sink, stop using them to communicate with your reports. https://lnkd.in/ehr_e2Xv

20

9 Comments

Posts of the week Engineering — Diffusion Models: https://lnkd.in/gbW2TvMp Git Cheat Sheet: https://lnkd.in/gaFKD99Y Google SRE Book: https://lnkd.in/ggj6ewDF Rootless Docker in a Multi-user Environment — It's All About Context: https://lnkd.in/gFY2WPfn AI — GoalChain is a simple but effective framework for enabling goal-orientated conversation flows for human-LLM and LLM-LLM interaction: https://lnkd.in/gibsrjcz EU Approves Historic AI Act: Pioneering Comprehensive AI Regulation: https://lnkd.in/gnCrHn8x Nice Reads — Minimal Temp Email: https://email.ml/ Don't Microservice, Do Module: https://lnkd.in/gRD7hq4G Mapping the Mind of a Large Language Model: https://lnkd.in/gSnh-sAw

1

LLMs and LangChain: The Dynamic Duo 🦸‍♀️🤖 Just finished a fun project where we built a Q&A bot using these two powerhouses. It's like having a personal AI assistant that's always ready to answer your burning questions! 🔥 Here's what I learned: LLMs are wizards: They can conjure up answers from thin air (or, uh, the internet). 🧙‍♂️ LangChain is the glue: It holds everything together and makes it easy to use. 🔗 Endless possibilities: This tech can be used for anything from customer support to writing catchy social media posts. 📝 Want to see it live (obvio a test server link 😛 ): https://lnkd.in/dZNAVCSV Note: it's a render link, they will take down the instance because of no use, you can hit the link and wait for two minutes, they will spin up the instance and you can try it out. Want to see the code? Check it out here: https://lnkd.in/dzjh4r47 #AI #MachineLearning #NLP #LangChain #LLM #Python #Developer #FunProject"

22

Interviewing at Google is an experience in itself. I shared my own interview experience from 2005 here - https://lnkd.in/gSVxtF45. "Challenge the status quo" is one of Google values which I experienced from the day I interviewed to this very day. I shared how this showed up in the article I mentioned above. Specifically what it means to be Googley or what is Googleyness (which in itself is an interview focus area). I also shared some interview tips in the article. How does your organization evaluate candidates? Do you evaluate for culture fit to create a homogeneous environment? Do you see if candidates would "challenge the status quo"? Share your thoughts and perspective.

39

6 Comments

Test time training of LLM. For easier problems providing context along with the prompt is enough. For harder problems like math, coding, law etc. taking a sample of similar problems and training few batches on it improves accuracy many X times (3-14 times or so). https://lnkd.in/gwp97RFr

1

During Prime Day 2018, significant parts of the Amazon Store were down for about 3 hours during the busiest shopping day of the year. It was a fascinating study in real-world systems architecture at scale: a sale item that was particularly popular caused a hot partition in a massive distributed hashtable, and this started a multi-level retry storm that brought many, many, many thousands of servers to their knees. I was here at the time, and I remember following the real-time investigation glued to my screen as details unveiled. If you want to know more, Jim Roskind has a really good reInvent talk where he gets into the technical details (https://lnkd.in/gbabs2dP). Amazon being Amazon, we took this failure as an opportunity to improve: rearchitected a lot of our systems, focusing on handling congestion, rethinking the way we retried failures, restricting unbounded queues, and, of course, load and stress testing. I was the Lead for Load and Performance Testing Infrastructure at Amazon at the time, so I found myself in the perfect middle of all this. It was a fascinating time!

40

4 Comments

𝙂𝘼𝙍𝘽𝘼𝙂𝙀 𝘾𝙊𝙇𝙇𝙀𝘾𝙏𝙄𝙊𝙉 - 𝙇𝙞𝙠𝙚 𝙩𝙝𝙚 𝙆𝙖𝙗𝙖𝙙𝙞𝙬𝙖𝙡𝙖! 🛢️ 🧠 𝐉𝐚𝐯𝐚 𝐁𝐮𝐟𝐟𝐞𝐭 : Day 1 of #30DaysOfJava Ever called the kabadiwala (scrap dealer) to clean out old newspapers and bottles piling up in your house? Every now and then, we need to get rid of things we don’t use anymore to free up space. In Java, this cleanup process is handled by something called Garbage Collection. Just like the kabadiwala goes around looking for items you don’t need, the Garbage Collector (GC) in Java hunts for objects in memory that aren’t being used by your program anymore. Once it finds them, it clears them out to free up memory, making sure your program runs smoothly without any memory clutter. Now, imagine calling the kabadiwala too frequently; he might start interrupting your routine. Similarly, if the Garbage Collector runs too often, it can slow down your application—a phenomenon called GC overhead. On the flip side, if he comes too late, your home will start overflowing with junk, just like memory issues in Java! 📌 Key Takeaway: Java’s Garbage Collector is like your friendly neighborhood kabadiwala, keeping your memory clean. But timing is everything—clean up too often, and you’ll slow down, wait too long, and you’ll run out of space! #java #oracle

21

5 Comments

Is cache invalidation one of the toughest and most complex problems a software engineer can face in his career? (rhetorical question) If you are interested in the problem, I suggest reading the following piece from the Netflix Technology Blog. #software #engineering #cache #invalidation https://lnkd.in/dSW6_-8M

4

Sharing a few of my recent Amazon shopping experiences here to confirm a few observations from people who may know. This is the third time in last few months when I was unable to return/replace the items I purchased. The product in question was a high ticket item and eligible for return/replacement. I was also able to place return orders for them smoothly from the Amazon shopping app. However, on the day of return pick up, I got an IVR call from Amazon saying the return pick didn't succeed because I wasn't present at pick-up/delivery address. In reality, I did not receive any phone call from any delivery agent and nobody came home to pick up return or deliver the replacement item. The call says pick-up/delivery will be attempted again twice and the same incident repeats the day after and the day after that. Now the order is forever stuck in return started/return failed state. And there's no way to mitigate this as Amazon app or website doesn't have the option for talking to a human customer service agent anymore. I think the delivery agents are well aware of this and they are exploiting it by not even attempting to deliver/pick up the items they were supposed to. Would like to know if anyone has experienced this and found a way to raise this issue and/or directly talk to a customer service agent. Appreciate your time and response. Thank you.

14

7 Comments

I've been at Amazon a long time. When I talk about my career with Amazon, the most frequent questions people ask me are: What kept me with Amazon for so many years? When I left after nearly 14 years, what drew me away? What brought me back 2 years later? The genuine answer to each, as trite as it sounds, is the same: it's the people The people who took a chance on me. The people that, when I struggled, helped me lean on my strengths and shore up my weaknesses. The people who saw potential and invested in me. The people who modeled the value of building relationships, not just building code. The people who challenged me, made me think differently, and helped me grow. As I've grown in my career it's also those people in whom I can see potential and invest, where I might take a chance to give them an opportunity, for whom my experience could be of some value as they grow, and with whom I can continue building lasting relationships. I've been with Amazon's Global Talent Management and Compensation for about a year and a half, where our mission is to help Amazonians build rewarding careers. In that time, I've met with hundreds of people in our org and every single one of them has been driven by our mission and passionate about how we can meaningfully improve the lives of every Amazon employee. I've learned so much about how challenging, rewarding, and important taking a culture-first approach can be. This post isn't a claim that Amazon is objectively unique, I know there are amazing people in every company, though it has been my personal experience that Amazon's intentionally peculiar culture is one that encourages these kinds of behavior in people. Rather, my point is this: To deliver sustained business and customer value, your organization cannot be a feature factory. It needs to be an organizational engine that continuously inspects and adjusts its culture to grow people and relationships, according to your values, and through which you deliver for your customers. #people #relationships #culture #amazon

263

8 Comments

https://lnkd.in/gqUkq3BM is an excellent overview of major topics in LLMs. If you are getting started in this space, this video which is also generated by another LLM 😂 , is a great place to begin your journey in this exciting area of AI. It covers topics like pre-training, evolution of GPTs and other popular LLMs, fine tuning (SFT, PEFT like LORA, RLHF), inference optimizations (quantization, speculative decoding, distillation, flash attention), output sampling techniques, multimodal capabilities and finally some real world applications like math, coding, medicine etc. Loved their analogies even for core complex concepts like flash attention - "it's like finetuning the engine of a car, to make it more fuel efficient, without changing how the car drives"

24