<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Hussein Karaki on Medium]]></title>
        <description><![CDATA[Stories by Hussein Karaki on Medium]]></description>
        <link>https://medium.com/@husseinkaraki?source=rss-17aae010d8c9------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*lp5TQ4Qgbjc-k5PXh9rqkQ.jpeg</url>
            <title>Stories by Hussein Karaki on Medium</title>
            <link>https://medium.com/@husseinkaraki?source=rss-17aae010d8c9------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 16:01:48 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@husseinkaraki/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Product Manage in Public: Day 3 Reflection — Stop Prematurely Optimizing (Everything)]]></title>
            <link>https://medium.com/@husseinkaraki/product-manage-in-public-day-3-reflection-stop-prematurely-optimizing-everything-571f4f7cb7f8?source=rss-17aae010d8c9------2</link>
            <guid isPermaLink="false">https://medium.com/p/571f4f7cb7f8</guid>
            <category><![CDATA[product-management]]></category>
            <dc:creator><![CDATA[Hussein Karaki]]></dc:creator>
            <pubDate>Tue, 30 Dec 2025 16:39:37 GMT</pubDate>
            <atom:updated>2025-12-30T16:39:37.244Z</atom:updated>
            <content:encoded><![CDATA[<h3><strong>Product Manage in Public: Day 3 Reflection — Stop Prematurely Optimizing (Everything)</strong></h3><p>I’m journaling my entire product management journey in public, on X, from ideation to shipped product. My mission is to bring awareness to product management practices to those building software products, especially in the startup scene, and learn some new things along the way.</p><p>I decided to start this series before I decided on what I was going to work on, because that would better properly capture the true intention of product discovery, in that the “what” is not nearly as important as the “why”.</p><p>It’s day 3. My brain is already a warzone. Notion pages, Apple Notes, Miro boards, physical sticky notes, journals and notebooks, half-baked thoughts everywhere…</p><p>This reflection is me trying to make sense of the mess and how I deal with it.</p><p>All real, all human, all me. Grok just cleaned up some awkward phrasing.</p><h3>Stop prematurely optimizing (Everything)</h3><p>As software engineers, we have a terminal disease: premature optimization. We do it with code, with workflows, with the we way we drive, with our entire damn lives.</p><p>Inefficiency spotted → perfect system imagined → google 17 alternatives → sign up for three tools → lose three hours → repeat forever.</p><p>I literally did it live while writing this. Seriously. Mid-sentence I thought:</p><p>“Wait. OoOo. Let me google Obsidian alternatives” (for the 18th time in 5 years, mind you)</p><p>“<a href="https://capacities.io/">Capacities.io</a>” Ok. Sign up.” (literally havent even tried the product yet)</p><p>We can’t help it. We love a mental challenge. And as long as you’re a software engineer, there’s little incentive to break the habit. Here’s the irony — <strong>premature optimization doesn’t scale</strong>. Time and mental energy are your scarcest resources. And in the world of startups, you’re signing up for “beginners stuck”.</p><p>So here’s the issue. Theres so much data, so much information, it overwhelms everyone. Setup creates the illusion of progress. I’ve personally struggled with this…starting projects but never finishing them. My github has over a 100 repositories with 1 commit each. This is inherently a psychological problem tied to our impulsiveness, dopamine addiction, and endless creativity. So lets create a new understanding for ourselves now:</p><p><strong>Optimization can happen when it hurts not to optimize.</strong></p><p>You’ll know it. It’ll be painfully obvious. So don’t worry about it ahead of time. The perfect folder structure, the ideal note-taking app, forget about it all. This is not laziness. <strong>This is psychological discipline. </strong>You have to learn to say “No” to yourself. You need to be able to tell yourself “chill the f* out cuz.” You need to actively <strong>capital M Manage</strong> yourself. The real problem isn’t the mess. <strong>It’s our intolerance of the mess. </strong>Wooosahhhh</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*z0PtSeZjrHKkP1Rhv-OsRg.png" /><figcaption>try it. you know you want to….woooooooo ……. sahhhhhhhh…..👂🤏 👁️👃🏻👁️ 👂🤏</figcaption></figure><h3>Be comfortable in chaos</h3><p>Different people have different tolerances for chaos and risk. That’s not a flaw. It’s a feature. The most successful people you admire didn’t get there by being perfectly organized from day one. They learned to operate effectively inside ever increasing levels of disorder and chaos. They trained themselves. Start by bringing awareness to it. That brings control to it. Now you have a choice:</p><ul><li>Ignore it (when it doesn’t matter)</li><li>Tolerate it (while you ship)</li><li>Manage it (when the cost of ignoring it gets too high)</li></ul><p>You can do this because:</p><ul><li>You refuse to prematurely optimize</li><li>You have psychological discipline</li><li>You can tolerate the mess</li><li>You can choose to manage yourself in chaos</li><li>You choose to embrace the risk, <strong>you badass mother fucker </strong>😤</li></ul><p>But here’s the big, uncomfortable twist (the startup / product-world yoga pose): when you enter into real product work — startup, growth-stage company, high-stakes corporate, you must become <strong>extremely</strong> good at <strong>organized chaos</strong>. Ruthless prioritization. Saying no to 97 things so the 3 things can actually move. Holding a thousand threads in your head without losing the most important one.</p><p>The spectrum is massive. Some people crumble under five open browser tabs. Others thrive with 50 sticky notes, three whiteboards, and a dozen half-finished notion pages. <strong>Your chaos tolerance is not fixed. It’s trainable.</strong> High performers thrive in this zone. So bring your awareness to it. The way successful product builders handle their mess today is very different from how most engineers handle theirs. Stop acting like your current mess is permanent. Start acting like the version of you that already thrives in organized chaos.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*I71p2l3W7Jh44mzGfxQqmQ.png" /><figcaption>This is ok. Organized chaos.</figcaption></figure><p>Key takeaways (save this for the next itch):</p><ul><li><strong>Don’t prematurely optimize </strong>— Let pain drive the refactoring.</li><li><strong>Manage yourself first </strong>— your impulse control translates to the real product you’re building</li><li><strong>Be comfortable in chaos</strong> — your personal ceiling is probably much higher than you think.</li><li><strong>Bring ruthless awareness </strong>to your own dopamine-seeking tendencies.</li><li><strong>Prioritize like your life depends on it </strong>— because in real product work, it kind of does.</li></ul><p>The tools will be there when you need them. Organization arrives when the cost of disorganization becomes unbearable. Until then: write the impulse down, close the fucking tab, and keep moving.</p><p>You’ve got this.</p><p>Join the “Product Manage in Public” community on X where I provide daily updates and journals as I go through the entire product management development process. DM’s are always open if you want to talk shop.</p><p><a href="https://x.com/hussein_builder">JavaScript is not available.</a></p><p>Thanks for reading!</p><p>Hussein Karaki</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=571f4f7cb7f8" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A Practical Agile Summary]]></title>
            <link>https://medium.com/@husseinkaraki/super-quick-agile-summary-3b1fcd7abd8e?source=rss-17aae010d8c9------2</link>
            <guid isPermaLink="false">https://medium.com/p/3b1fcd7abd8e</guid>
            <category><![CDATA[waterfall]]></category>
            <category><![CDATA[agile]]></category>
            <dc:creator><![CDATA[Hussein Karaki]]></dc:creator>
            <pubDate>Fri, 12 Jul 2024 23:07:53 GMT</pubDate>
            <atom:updated>2024-07-16T17:09:15.291Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/590/1*Tb9HE4K__Hbfy0IDnsWvBA.png" /><figcaption>Over simplistic analogy for Agile vs Waterfall</figcaption></figure><p>Hey guys. So a new grad asked me about Agile vs Waterfall today. I thought I would share with her my personal take on it. To summarize, the most important distinction between these two software methodologies is deployment cycle time. Waterfall has a deadline; its main objective is to complete the full software product with all its features by the deadline. This approach was problematic for several reasons.</p><p>Firstly, it required developers to estimate how much work it would take to build out all the features. It turned out to be very difficult to do that accurately, and more often than not, they did not allocate enough time for the work.</p><p>Another issue was that the customer would often think they wanted a set of features but, in reality, would not actually use them. It’s challenging for a business to communicate its requirements and needs, especially across other businesses and organizations. This problem becomes even more apparent as the size of the application grows.</p><p>Another problem is the amount of effort developers put into the project at the beginning vs. the end. Naturally, people tend to procrastinate, especially when dealing with bigger tasks. This is true on an organizational level too. The trend of effort and difficulty tends to be flat initially and then climbs quickly towards the second half to the last quarter of the project timeline. This is problematic for obvious reasons, especially when combined with the issue of underestimation.</p><p>Additionally, Waterfall methodologies, given their big releases, tend to have bigger “monolith” architectures as opposed to microservice architectures, with many developers working on one big app. Now, this does not necessarily have to be true, but tends to be the case. This association with monolithic architectures is particularly problematic in today’s fast-paced development environment, where flexibility and scalability are crucial.</p><p>This led to another set of challenges regarding testing. Testing all the pass / fail cases of each individual component is significantly easier than testing how a pass / fail propagates through a monolith. Microservices, when combined with, for example, RESTful standards, influences us to use respective HTTP codes to reflect the result of that transaction. SDK’s come with built in support for recognition, detection, and handling for HTTP 4xx and 5xx codes. Although developers partly need to explicitly handle these codes, they’re already on a good starting off point. Then there is the issue of managing and understanding the whole app and its features. Onboarding new people onto the project becomes significantly more difficult. The list goes on and on.</p><p>With the rise of Agile methodologies, we’ve seen a natural shift towards microservices and product-based environments. This isn’t just coincidence — Agile principles actively encourage modular, iterative development that aligns well with microservice architectures. For example, companies like Netflix and Amazon have famously adopted both Agile and microservices to great success, demonstrating how these approaches complement each other in practice.</p><p>Agile solved a lot of problems because it forced developers to demo their progress every two weeks. This allowed them to course-correct early on and get regular, consistent feedback from the customer. The customer could say, “Yes, we like this feature, continue working on it, maybe even expand it,” or “No, we don’t like this feature, let’s stop developing it now and minimize our losses.”</p><p>Agile also encouraged a microservice architecture. Smaller services are easier to maintain, and tech debt is easier to control. Tech debt, which refers to the implied cost of additional rework caused by choosing an easy solution now instead of using a better approach that would take longer, is such an issue that we’re not even sure we can handle it, despite numerous rules, patterns, tools, and processes to minimize it. We do know one thing for sure: because microservices have less functionality per code and service, they create less tech debt — a matter of ratio vs. size. We’re able to compartmentalize it better with microservices.</p><p>There are also challenges with microservices, such as chattiness (excessive communication between services) and cost, but generally speaking, for long development of a sophisticated complex application, microservices are generally objectively better. This view is supported by numerous case studies and industry reports. For instance, a 2020 O’Reilly survey found that 77% of organizations had adopted microservices, with 92% experiencing success with this architecture.</p><p>While some critics argue that Agile can lead to scope creep or lack of long-term planning, these issues can be mitigated through proper implementation of Agile practices such as regular backlog grooming and maintaining a product roadmap. The benefits of Agile — including faster time-to-market, better alignment with customer needs, and improved ability to adapt to changes — far outweigh these potential drawbacks in most scenarios.</p><p>Note: Mostly written in my own words, this summary was lightly edited for clarity and readability with the assistance of Claude, an AI language model.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3b1fcd7abd8e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Spring Boot + Kafka on Azure Event Hubs]]></title>
            <link>https://medium.com/@husseinkaraki/spring-boot-kafka-on-azure-event-hubs-a20009f4bb78?source=rss-17aae010d8c9------2</link>
            <guid isPermaLink="false">https://medium.com/p/a20009f4bb78</guid>
            <category><![CDATA[spring-boot]]></category>
            <category><![CDATA[kafka]]></category>
            <category><![CDATA[java]]></category>
            <category><![CDATA[azure-event-hub]]></category>
            <category><![CDATA[azure-kafka]]></category>
            <dc:creator><![CDATA[Hussein Karaki]]></dc:creator>
            <pubDate>Mon, 29 Apr 2019 13:31:00 GMT</pubDate>
            <atom:updated>2019-04-29T19:17:47.904Z</atom:updated>
            <content:encoded><![CDATA[<h3>A Quick Guide to Spring Boot + Kafka on Azure Event Hubs</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*h3EbXDPPxsRn-GGq7opWdQ.jpeg" /><figcaption>photo cred: <a href="https://www.pexels.com/@ashutoshsonwani">https://www.pexels.com/@ashutoshsonwani</a></figcaption></figure><h3>Intro</h3><p>This is a quick guide on how to implement Kafka in Spring Boot against Azure’s Event Hubs. I found the tutorials online did not provide exactly what I was looking for, and after we finished implementing Kafka for some of our new services, my colleague Nick recommended I write an article on the topic…(ahem). Without further ado, here is a quick guide on how to send/receive java objects as JSON to and from Azure Event Hubs. I am in no way an expert, nor do I think this is the best way to do this, so advise, recommendations, and constructive criticism are welcome and encouraged.</p><p>We’re deploying our services on PCF, and using Azure’s Event Hubs service as our pub sub infrastructure. <a href="https://docs.microsoft.com/en-us/azure/event-hubs/event-hubs-for-kafka-ecosystem-overview">Azure offers a kafka interface on top of Event Hubs</a>. Topics, partitions, and consumer groups are all supported. Microsoft does provide a SDK, but we’re going to ignore it to avoid potentially legal overhead and bureaucratic costs with implementing a 3rd party library or SDK within an organization.</p><h3>Special Cases</h3><p>When it comes to sending data over kafka, there are lots of things to consider. We won’t go into detail here, but understand that you may want to use a different implementation depending on your use case. For example:</p><ul><li>You don’t have control of both the producer and consumer</li><li>If there are multiple schemas being sent to the same topic</li><li>Minimizing latency is absolutely critical</li></ul><p>Covering those cases is outside the scope of this guide. If you would like more information on how to handle those cases, I recommend visiting <a href="https://docs.spring.io/spring-kafka/reference/html/#kafka">Spring Reference</a>, <a href="https://docs.spring.io/spring-boot/docs/current/reference/html/boot-features-messaging.html#boot-features-kafka">Spring Docs</a>, &amp; <a href="https://gitter.im/spring-projects/spring-kafka">Spring Kafka’s Gitter</a>.</p><h3>Setup</h3><ol><li>Setup your Azure Event Hubs. <a href="http://[https://docs.microsoft.com/en-us/azure/event-hubs/event-hubs-create-kafka-enabled">Here is a tutorial</a> on how to setup up a kafka-interfaced Event Hubs. Store the “Connection String-primary key” somewhere safe, we will use it later.</li><li>Add dependencies to your spring boot app (and don’t forget to include <a href="https://projectlombok.org/mavenrepo/">lombok </a>:D ).</li></ol><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/023143e3d833ac1287eac32cb806c4fa/href">https://medium.com/media/023143e3d833ac1287eac32cb806c4fa/href</a></iframe><h3>Implementing the Producer</h3><p>The spring framework kafka library includes a built in Json serializer and deserializer. You can send and receive entire java objects by first serializing and deserializing the java objects to and from JSON, and configuring the serializer and deserializer into the kafka producer and consumer.</p><p>Lets first start with the producer. You’ll need to inject the pre-configured KafkaTemplate, so you can publish the messages.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/fe250cc302cc1ae5107f4c0525ec267b/href">https://medium.com/media/fe250cc302cc1ae5107f4c0525ec267b/href</a></iframe><p>And your <em>producer </em>data transfer object. This object will be serialized to json by the configured serializer.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/84e9e2b75bbc14b47b4107b0b59eadf5/href">https://medium.com/media/84e9e2b75bbc14b47b4107b0b59eadf5/href</a></iframe><p>Kafka provides serializer and deserializer implementations. You’ll find the JsonSerializer &amp; JsonDeserializer in org.springframework.kafka.support.serializer. Extending the class like so allows you to reference the Serializer in configuration while providing a type to deserialize the object to. I found this was the only way I could reference the class from configuration while still specifying the input type requirement. If you decide to programmatically configure your publisher, this may not be necessary as you can just instantiate a new JsonSerializer&lt;&gt;(ProducerMessage).</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/d02c3a7b9ee743706d3c83117cb7d7b6/href">https://medium.com/media/d02c3a7b9ee743706d3c83117cb7d7b6/href</a></iframe><p>And your producer configuration. bootstrap-servers is the url of the kafka queue. client-id isn’t necessary, but helps identity what services are doing the publishing and subscribing. The properties under kafka are specific for Azure Event Hubs and are required. You have flexibility here — you can programmatically configure your beans if you’d like by creating them yourself, specifically the prducerConfigs, ProducerFactory, and KafkaTemplate. Look for KafkaAutoConfiguration in org.springframework.boot.autoconfigure.kafka for reference.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/245cfe9774212641d20978b4bc323bbb/href">https://medium.com/media/245cfe9774212641d20978b4bc323bbb/href</a></iframe><h3>Implementing the Consumer</h3><p>Your consumer. This is where you’d make your service call.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/5b0b14cd320f54a86b130c242a701f98/href">https://medium.com/media/5b0b14cd320f54a86b130c242a701f98/href</a></iframe><p>Your consumer dto. Make sure you have a @NoArgsConstructor, as the deserializer will be creating an instance of the object.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/61889a831831be3fb948277acf9f758d/href">https://medium.com/media/61889a831831be3fb948277acf9f758d/href</a></iframe><p>And the consumer configuration. Similar to the producer configuration, you can set these programatically. By defualt, the built in deserializer requires you declare the original package the json was serialized from. This creates tight coupling. The spring.json.use.type.headers config allows you to deserialize without explicitly declaring the json as an object from a specific package. The spring.json.value.default.type config is the java type the json deserializes to.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/ddf42997cebace891a9e4fe9bc7154af/href">https://medium.com/media/ddf42997cebace891a9e4fe9bc7154af/href</a></iframe><h3>Testing</h3><p>Write a test controller, call it via Postman, and check your consumer service for the log.</p><p>And that is it :)</p><p>Tweet me at @KarakiDev</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a20009f4bb78" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>