Burak Serdar Özşimşek

https://lnkd.in/ePnJZw97 A large-scale leak of information has been recorded, which includes more than 183 million e-mail passwords. Tens of millions of passwords fell on the account of "Gmail". Cybersecurity experts said that the incident is one of the largest incidents of this kind. Oxu.Az it is reported that the publication" New York Post " (NYP) wrote about this. The total volume of leaked data amounted to 3.5 terabytes. The information was collected for a year as a result of the activities of malware that secretly stole account information from users ' infected devices. Experts say that most of the information was revealed in other cases of leaks, but the relevance of millions of "Gmail" passwords has been confirmed. The leak affected not only the users of" Gmail", but also the users of" Outlook"," Yahoo " and hundreds of other web platforms. Experts emphasize that this incident is not the result of hacking "Google" servers, but is associated with the theft of data from user computers through virus software. Leaked passwords also give villains access to other services of their victims during the overlapping of login and password pairs. Experts recommend changing passwords immediately, enabling two-factor authentication, and using password managers as a protective measure. Special online services can be used to check the presence or absence of account information at the time of the leak.

2

Even after 25 years in this industry - whether you're a Contractor or Employee - you still participate in case studies during interviews. It's a natural part of the process. We also gave case studies when hiring for our team. In my opinion, this method is a great opportunity to see a person's perspective, coding skills, attention to detail, architectural approach, and soft skills. Even though we've entered the AI Age, some values remain unchanged: "Actions speak louder than words." Case studies are the perfect example of this. I'm in a case study process with a company I won't name. I built this task for them but will keep it on my GitHub as a reference. Although I've spent a long time in the .NET ecosystem, I also had Java knowledge from my past. What I appreciated about this company was their focus on engineering disciplines rather than technology fanaticism. They didn't hang up saying "Oh .NET? No, we're looking for Java." They asked "Would you be comfortable?" I said "Languages are just tools." And they truly are. I also used AI in this project. But I used AI as a guide: to apply disciplines I already know, instead of searching on Google, to speed up testing and development processes. In short, I accelerated my work and improved its quality. Lately, we see many posts like "If you write code with AI this way you'll fail, that way something else happens." It reminds me of those who said "This is the devil's work" when the printing press was invented. Don't resist innovation. Embrace it, examine it, improve it. How others use it doesn't concern you - what concerns you is how YOU use it. Open-mindedness becomes valuable here. While building this case, I also thought about the reviewer. We wrote scripts to make their job easier with AI - I dictated, it wrote. Then I couldn't stop: I turned Playwright-based E2E tests into a presentation with "make present" command. Added some jokes for fun and this is what came out 🎬  ☕ If you want to see details: git clone → make start → make present → sip your coffee  📦 https://lnkd.in/dutZ78zu  #java #springboot #microservices #kafka #docker #casestudy #softwareengineering #career 🎬 https://lnkd.in/deMNEcSz

5

2 Yorum

#DEVFEST2025: 1M RPM for 100M... Architecture of High Performance ML Solutions ⚙️ Delivering machine learning at scale requires more than just powerful models—it demands high-performance systems that can serve predictions reliably under massive load. In this talk, we will walk through the architecture behind Trendyol’s ML solutions that power real-time experiences for over 100 million customers, handling up to 1M requests per minute. We’ll explore how we leverage Google Kubernetes Engine (GKE), Kubernetes, gRPC, and NVIDIA Triton Inference Server, alongside our internal deployment and management tools, to achieve low-latency inference at scale. 🧩 Omar Elcircevi is a Machine Learning Engineer at Trendyol Group, where he builds high-performance ML platforms and automated solutions that serve millions of customers at scale. He has delivered projects spanning data analytics, deep learning models, and generative AI applications. Beyond his engineering role, Omar is also an active organizer of Google Developer events in Turkey. ❗ Join us to gain practical insights into building resilient, efficient, and production-ready ML infrastructures that balance speed, scalability, and cost!! 🚀 TICKETS >> https://lnkd.in/eGsfQUsX ⚡Don't miss the chance to learn from an industry expert !! 🕐 Date: 22 Nov 2025 📍 Venue: ADA University

20

A travel and tourist company in Turkey relies on Azure for storage and storage-related products. According to a product director there, the company achieved a 40% cost advantage with Azure, which it uses to gain a 30% efficiency back into their system.

3

A travel and tourist company in Turkey relies on Azure for storage and storage-related products. According to a product director there, the company achieved a 40% cost advantage with Azure, which it uses to gain a 30% efficiency back into their system.

2

Turkey's e-commerce leader, Trendyol Group, leveled up by migrating from relational databases to #NoSQL with #Couchbase 👏 Learn how they built a scalable, resilient system in this blog by Software Engineer Emre Tanrıverdi: https://bit.ly/3WSSo5V Emre Savcı

16

Az once fantastik uzunlukta bir patch serisi geldi. https://lnkd.in/dqQFWZja Linux support for a standalone FIPS 140 module                                                                                            Hi all,                                                                                                                                                                                        This patch set adds support for building and loading a standalone FIPS 140 module. This is mostly useful for distributions that want to certify their kernel's crypto code with NIST. Please see Documentation/crypto/fips140.rst for more details.                                                                                               I apologize for the large patch series. I could have squashed it down to fewer commits but it would really make it harder to see what's going on.                                                                                               A note on terminology, since this is a source of confusion: a "FIPS  module" in NIST jargon means the binary blob submitted for FIPS certification (i.e. typically the entire kernel), and has nothing to do with kernel modules per se. However, in this proposal, the "FIPS module" does actually also correspond to a singular kernel module, fips140.ko.                                                                                                                                                         FIPS certifications validate a specific binary, not the source code used to build it. Currently, most if not all distros submit the entire kernel for certification, and the binary certification is invalidated if any part of the kernel changes -- even if those changes are unrelated to cryptography. Fortunately the FIPS process allows for vendors to "attest" that the changes don't affect cryptographic operations within the kernel, but the fact remains that the binary certification is technically obsolete as soon as any bug or security fix is applied to the kernel.                                                          This proposal attempts to separate the crypto part of the FIPS module  from the rest of the kernel. This allows us to freeze the binary FIPS module and reuse it in newer versions of the kernel without invalidating the certificate on the FIPS module. 😎

4

MCP və təhlükəsizlik. Yazılanları xarici bir linkedin üzvünün postundan götürdüm: MCP security just became everyone's problem ChatGPT drops MCP support Wednesday. By Friday, someone's already weaponizing calendar invites to steal emails. ;) Here's the attack: 1) Send victim a calendar invite with hidden prompt injection 2) Victim asks ChatGPT "check my calendar for today" 3) ChatGPT reads the malicious invite and gets completely hijacked 4) Now it's dumping private emails, sending everything to the attacker Bu təhlükəsizlik məsələlərinə görə MCP indiyədək işlətməmişəm. İstər coding alətləri, istər MCP-də bilmədiyimiz boşluqlar var. Nəzərinizdə olsun.

9

The OWASP Top 10 keeps evolving and so do attackers. Are you keeping up? In this hands-on training, you’ll dive into the latest OWASP Top 10 risks and learn how to identify and test them in practice using Burp Suite. No fluffy theory, just the core principles of web application security and how to actually validate them. Whether you’re: 🔨 A builder (developer) 💥 A breaker (tester or red teamer) 🛡️ A defender (DevOps and SRE or blue team) This course helps you understand vulnerabilities from every angle and gives you the skills to spot them before someone else does. Are you ready to think like an attacker and act like a defender? Then sign up for this training 👉 https://lnkd.in/esMcKuJC

10

Bir çox mütəxəssislər tez-tez Agile və Scrum anlayışlarını qarışdırırlar. Əslində isə bunlar eyni şey deyil. 🌀 Agile — bu bir fəlsəfədir, düşüncə tərzidir, Agile Manifestoda qeyd olunan dəyərlər və prinsiplər toplusudur. Onun əsas məqsədi müştəri üçün dəyər yaratmaq, dəyişikliklərə çevik uyğunlaşmaq və əməkdaşlığı ön planda saxlamaqdır. 🧩 Scrum isə Agile prinsiplərini praktikada həyata keçirmək üçün yaradılmış bir çərçivədir (framework). O, konkret rollar (Scrum Master, Product Owner, Developers), artefaktlar (Backlog, Increment və s.) və hadisələr (Sprint, Review, Retrospective) təqdim edir ki, proses şəffaf və səmərəli olsun. Sadə dillə desək: 👉 Agile — “niyə” 👉 Scrum — “necə” Yəni, yalnız Scrum qaydalarını bilmək kifayət deyil — əsas məsələ Agile düşüncə tərzini anlamaqdır. Many professionals often confuse Agile and Scrum. In reality, they are not the same thing. 🌀 Agile is a philosophy — a mindset based on values and principles described in the Agile Manifesto. It’s about delivering value step by step, adapting to change quickly, and collaborating closely with the customer. 🧩 Scrum is one of the frameworks that helps teams put Agile principles into practice. It defines specific roles (Scrum Master, Product Owner, Developers), artifacts (Backlog, Increment, etc.), and events (Sprint, Review, Retrospective) to make the process transparent and efficient. Simply put: 👉 Agile is the “why” 👉 Scrum is the “how” That’s why understanding the Agile mindset is more important than just following Scrum rules. #Agile #Scrum #AgileAzerbaijan #ScrumMaster #AgileDüşüncə #Komandaİşi #Layihəİdarəçiliyi #DəyişikliklərəUyğunlaşma #AgileCoach #AgileLiderlik #ProductOwner #AgileMindset #Leadership #ProjectManagement #Teamwork #ContinuousLearning #Adaptability

10

This was just one of the real-world challenges we tackled during our intensive 4-day Elasticsearch training with HAVELSAN this week! Over 4 days, we went deep into: ✅ Cluster architecture & management best practices ✅ Advanced data modeling & aggregations ✅ Search optimization & analyzer configuration ✅ Ingest pipelines & data transformation ✅ Heap management & performance tuning (including the critical compressed oops threshold) ✅ Operations, monitoring & troubleshooting at scale The highlight? Watching the team navigate through heap sizing challenges, understanding why "bigger isn't always better" when it comes to JVM memory, and implementing real-world solutions for cluster stability. Huge thanks to the HAVELSAN team for their engagement and excellent questions that made this training so dynamic! 💪 Are you facing Elasticsearch challenges in production? What's been your biggest pain point? #Elasticsearch #DataEngineering #BigData #Search #HAVELSAN #Training #Performance #DevOps

4

DataCube Bootcamp: Python-da Data Strukturlarının Gücü! Bu həftə DataCube Bootcamp-də mövzumuz Python-un təməl sütunları olan Data Structures (Məlumat Strukturları) idi. Öyrəndiklərim sadəcə sintaksis deyil, məlumatı daha səmərəli idarə etməyin yolları oldu. 🔍 Nələri öyrəndik? Əsas növlər: Integer, Float, String və Boolean. String Metodları: split(), join(), replace() kimi metodlarla mətnləri formatlamaq. Kolleksiyalar: * List: Siyahıları append() və sort() ilə idarə etmək. Tuple: Məlumatın dəyişməzliyini (immutable) təmin etmək. Set: Təkrarlanan dəyərləri təmizləmək və kəsişmələri tapmaq. Dictionary: Açar-dəyər cütlükləri ilə sürətli məlumat əldə etmək. 💡 Əsas nəticə: Python-da data növlərini bilmək — məlumatları düzgün təşkil etmək və effektiv kod yazmaq bacarığıdır. Dərsi maraqlı və anlaşılan şəkildə izah etdiyi üçün @HüseynAbdullayev təşəkkürlər! #Python #DataCube #Bootcamp #DataStructures #Programming #Learning 🚀 DataCube Bootcamp: The Power of Data Structures in Python! This week at DataCube Bootcamp, we focused on the core pillars of Python: Data Structures. I’ve learned that it’s not just about syntax, but about how to organize and manipulate data effectively. 🔍 Key Takeaways: Basic Types: Deep dive into Integer, Float, String, and Boolean. String Methods: Formatting and manipulating text using split(), join(), and replace(). Collections: * List: Managing dynamic data with append() and sort(). Tuple: Ensuring data integrity through immutability. Set: Removing duplicates and using operations like intersection() and union(). Dictionary: Efficient data retrieval using key-value pairs. 💡 Main Insight: Mastering data structures is the key to organizing information and writing clean, efficient code. Big thanks to Hüseyn Abdullayev for making these topics so engaging and easy to understand! #Python #DataCube #Bootcamp #Coding #DataScience #TechEducation

21

🚨 Your AI passes every test… until reality changes. What happens when data shifts? Pipelines evolve? Outputs behave unpredictably? At Test Automation Summit – Istanbul 2026, Alper Donmez, VP of Quality Assurance at Jotform and a Featured Speaker, dives into the real question: 👉 Can you actually trust your AI before production? 🎤 Session: How to Create an Environment to Test Your Own AI 🔎 What makes this session unmissable? ✔ Testing beyond code—datasets & pipelines ✔ Measuring AI behavior with real observability ✔ Building reproducible, reliable systems Because in AI, passing tests isn’t enough. 📍 Istanbul 📅 31 March 2026 🎟️ https://lnkd.in/gswdcifR ⏳ Don’t just test AI. Learn how to trust it. #AItesting #TestAutomation #QualityEngineering #SoftwareTesting

3

1 Yorum