<?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 Lukas Jan Stroemsdoerfer on Medium]]></title>
        <description><![CDATA[Stories by Lukas Jan Stroemsdoerfer on Medium]]></description>
        <link>https://medium.com/@ljstroemsdoerfer?source=rss-1bddc0d22949------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*dtFQF8upNy4yFQ7DW4yfQg.jpeg</url>
            <title>Stories by Lukas Jan Stroemsdoerfer on Medium</title>
            <link>https://medium.com/@ljstroemsdoerfer?source=rss-1bddc0d22949------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 12:50:42 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@ljstroemsdoerfer/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[The good, the bad and the ugly of becoming a Data Science Leader]]></title>
            <link>https://medium.com/@ljstroemsdoerfer/the-good-the-bad-and-the-ugly-of-becoming-a-data-science-leader-e3171ced94f1?source=rss-1bddc0d22949------2</link>
            <guid isPermaLink="false">https://medium.com/p/e3171ced94f1</guid>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Lukas Jan Stroemsdoerfer]]></dc:creator>
            <pubDate>Sun, 20 Feb 2022 18:14:30 GMT</pubDate>
            <atom:updated>2022-02-20T18:14:30.965Z</atom:updated>
            <content:encoded><![CDATA[<h4>What changed to the better, to the worse and what surprised me, when moving from a Data Scientist to a Leadership position</h4><p>I took over my first Data Science leadership role last September. Before that, I spent a couple of years as a ML engineer and Data Scientist working on various projects and products. After roughly 100 days in my new role, I took some time to reflect and I want to share that with you.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*scBRPp0dV40ThAUv" /><figcaption>Photo by <a href="https://unsplash.com/@sewellkare12?utm_source=medium&amp;utm_medium=referral">Karen Sewell</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h3>What changed to the better</h3><p>In my previous role I could create value by solving complex problems and establishing new ways of doing things. That was great and I loved it. In my new role, the direct value creation by solving problems is rather secondary. I spent most of my time helping my team to stay sharp, get even better and deliver kick-ass results when killing it in their projects.</p><p>I do that in three distinct ways: (i) positioning our efforts strategically, so that all our projects have the necessary funding, attention and also the recognition; (ii) spending a lot of time to understand what my team mates need, want and how to best spark their creativity; (iii) defining best practices, templates, and giving advice on architectural decisions.</p><p>This is extremely rewarding and I love (nearly) every aspect of it. In my new role, I am able to shape things for the better and establish Data Science in our company and ensure it’s long-term success. Previously, I my impact on this front was limited. So that definitely changed for the better and it keeps me going.</p><h3>What changed for the worse</h3><p>Well, I kind of expected it, but it still bugs me every day. I love getting “in the zone” and coding for hours straight. Naturally, in my new role, there is hardly time for that. If I do get some coding time, it is usually after hours. I have to get used to that, but is proving difficult.</p><p>The second aspect that changed for the worse, is definitely my struggles with relaxing after work. I am working on this and it is getting better. However, the heterogeneity of topics I am handling daily is so massive, that it kind of overwhelmed me at the beginning.</p><h3>What really surprised me</h3><p>Looking at my main responsibilities, I was sure that defining best practices and stuff is gonna be a lot of fun and I will love it. Working with my team and investing in my people was something I was looking forward to a lot and I absolutely love it.</p><p>What genuinely surprised me though, is that the marketing and campaigning, I have to do to make our effort visible, is also super fun. That was the part I dreaded the most when moving into my current position. Surprisingly it is fun. I created a pitch deck and a few demo cases and with those I am making my rounds.</p><h3>Conclusion</h3><p>After 100 days in my Data Science Leadership role, I am still happy I made the transition, however, I need to learn to live with the consequences. The stress that comes from the variety of topics is something I am trying to handle with really detailed planning on my side. I guess the coding part, I will have to accept. Though, I missing the coding, it is extremely rewarding to see the positive impact I can have on my team and on Data Science in my company in general.</p><p>To sum it up: great stuff!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e3171ced94f1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Impress with your application]]></title>
            <link>https://medium.com/@ljstroemsdoerfer/impress-with-your-application-c4674c026f5?source=rss-1bddc0d22949------2</link>
            <guid isPermaLink="false">https://medium.com/p/c4674c026f5</guid>
            <category><![CDATA[careers]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Lukas Jan Stroemsdoerfer]]></dc:creator>
            <pubDate>Tue, 28 Dec 2021 21:58:40 GMT</pubDate>
            <atom:updated>2021-12-29T08:42:41.527Z</atom:updated>
            <content:encoded><![CDATA[<h4>Building a kick-ass Data Science application to stand out and get the job</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*iwZnCVMxrvI1DwNC" /><figcaption>Photo by <a href="https://unsplash.com/@mkvandergriff?utm_source=medium&amp;utm_medium=referral">Maranda Vandergriff</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><p>Writing applications is often boring and tedious, which hinders creativity. As a result hiring managers get tons of look-a-like applications. To increase your changes of getting job, a creative and authentic application can help.</p><p>I had a bigger interview for my dream position and an extremely important next career step. So I decided to polish my application resources and build something that lets me stand out. I put all those resources into a Git repo, you can easily recreate what I am doing here:</p><p><a href="https://github.com/LJStroemsdoerfer/applyiverse">GitHub - LJStroemsdoerfer/applyiverse: Everything you need to apply for Data Science Positions</a></p><h4>The CV &amp; Cover Letter</h4><p>CVs and cover letters leave quite a lot of room for creativity. I build mine with LaTeX. There are thousands of cool templates out there and I happened to stumple upon the <a href="https://www.latextemplates.com/template/developer-cv&quot;">developer cv</a> from Jan Küster and Jan Vorisek. I tweaked the layout a bit and ended up with a dark themed CV and letter that I really liked.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/678/1*UyVmkv7AVljlixtLi7Raig.png" /></figure><p>The cover letter follows a similar design. You can of course adjust everything to your liking.</p><h4>The Interview</h4><p>More often than not you will have to deal with more than typical Q&amp;As during your interview. More and more hiring managers — especially for Data Science — are relying on coding tests, showcase, case studies and what not. To present such tasks, I developed a little website, which you can freely customize to your liking. Why a website you ask? Well, honestly, I am not a big fan of PowerPoint and you need some presentation material. To me it feels more authentic and natural to do it that way.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6XKjC_tzVMgJKj3sBwLOHA.png" /></figure><p>The website is based on mkdocs, it’s a standard documentation framework for python and I absolutely love it. It’s fast to build and it looks amazing. So I thought, what the heck, and I build the website on top of it.</p><p>The neat thing about mkdocs is, that you can include a Jupyter notebook. For the showcase, or case study, whatever you have, a Jupyter notebook is a nice way to present your results. In the repo, I build a small showcase to predict tomorrow’s Bitcoin high in euros. Feel free to adjust it, but you can also use it of course.</p><h4>Conclusion</h4><p>Kick-ass creative applications are definitely going to boost your chances of getting invited and showing off with a nice little website is for sure gonna impress.</p><p>All resources described here are free to use and you can find those in the repository I linked in the beginning.</p><p>I am looking forward to your feedback and happy applying 😉</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=c4674c026f5" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What I learned leading a Data Science Team]]></title>
            <link>https://medium.com/data-science/what-i-learned-leading-a-data-science-team-e8ac1f107a89?source=rss-1bddc0d22949------2</link>
            <guid isPermaLink="false">https://medium.com/p/e8ac1f107a89</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[leadership]]></category>
            <dc:creator><![CDATA[Lukas Jan Stroemsdoerfer]]></dc:creator>
            <pubDate>Tue, 14 Dec 2021 07:49:52 GMT</pubDate>
            <atom:updated>2021-12-15T15:10:53.014Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*gTSSqgdcW-rBqRO-" /><figcaption>Photo by <a href="https://unsplash.com/@nasa?utm_source=medium&amp;utm_medium=referral">NASA</a> on <a href="https://unsplash.com/?utm_source=medium&amp;utm_medium=referral">Unsplash</a></figcaption></figure><h4>Transitioning from a Data Scientist role to a management position is more difficult than I thought, but I am learning every day</h4><p>When I took on the leadership of a Data Science team, I had little people management experience. As a Data Scientist myself it was easy to connect, but soon I got to face some challenges keeping the team happy and on track.</p><p>What follows is a list of my learnings. Are they exclusive to Data Science teams? Probably not, but then again, I couldn’t tell.</p><h4>1. Enterprise IT can be a tough environment</h4><p>This wasn’t really big news to me as I worked on multiple Data Science projects in enterprise environments before, however, I had never considered the effect on long-term happiness. Don’t get me wrong, enterprise IT has vastly different regulations and precautions to comply to than most Data Science products. However, rigid — even if rightfully so — processes tremendously slow down the thought and solution process in Data Science. This will lead to frustration and hinder creativity at some point.</p><p>How to deal with that? Well, I haven’t found a one-fits all solution. For us having an enterprise compliant development space for Data Scientists and a ton of alignment meetings to early on involve enterprise IT experts is helping us.</p><h4>2. Buried Data Scientists tend to leave</h4><p>This I had to experience the hard way. Many enterprise projects are long-term oriented and they have to be. Having a Data Scientist from the beginning to the end in such a project will have some unwanted side effect apart from the desired continuity in the project team. An end-to-end business application with Machine Learning at its core is difficult to build, yet the Data Science related work makes only part of it. So, the Data Scientist will most surely end up taking on a bunch of task not really related to her or his profile. This is okay for some time, but as soon as it becomes the new normal, Data Scientists will start feeling unhappy.</p><p>What seems to work for us, is to have Data Scientists work on multiple projects at the same time and commit to a certain workload before each iteration. This way, mostly profile related tasks are agreed upon.</p><h4>3. Remote Work is an absolute Game changer</h4><p>Somehow, most of our team loves working remotely — including myself — and why wouldn’t they, it’s amazing. Sure, it has its drawbacks, but our productivity skyrocketed and honestly, somehow with all the Git based discussions, tech meetings and 1:1 meetings, we developed a new kind of intimacy in the team.</p><p>What I appreciate even more is that we opened up about our personal lives more. The countless times my little one greeted my colleagues or a dog complained about something, brought us together.</p><p>I am fully aware of the negative side effects here, however I try to focus on the positive.</p><h4>Conclusion</h4><p>Most experienced Data Scientists are amongst the most intrinsically motivated people. There are now and then formal education paths popping up, however most of us come from various disciplines and started to get into Data Science out of pure interest. With a lot of dedication we learned and grew as Data Scientists. To keep Data Scientists happy, we need to create a fitting environment.</p><p>Why do I share this? I am guessing some of you are already pretty experienced leaders in the Data &amp; Analytics space. However, like me, there are probably many other Data Scientists starting to grow into leadership position now. I would love to hear your thoughts on this, so reach out :)</p><p>PS:</p><p>The hardest learning isn’t anything team related. It’s how hard it is for me to let coding go. I am still coding, but way less than before. I am okay with this, cause there are other ways for me to create value and contribute, but it’s definitely a learning. And not an easy one.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e8ac1f107a89" width="1" height="1" alt=""><hr><p><a href="https://medium.com/data-science/what-i-learned-leading-a-data-science-team-e8ac1f107a89">What I learned leading a Data Science Team</a> was originally published in <a href="https://medium.com/data-science">TDS Archive</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Machine Learning deployment with Python]]></title>
            <link>https://medium.com/data-science/machine-learning-deployment-with-python-e64a37c3155a?source=rss-1bddc0d22949------2</link>
            <guid isPermaLink="false">https://medium.com/p/e64a37c3155a</guid>
            <category><![CDATA[docker]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[kubernetes]]></category>
            <category><![CDATA[deployment]]></category>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Lukas Jan Stroemsdoerfer]]></dc:creator>
            <pubDate>Sun, 07 Jun 2020 11:35:34 GMT</pubDate>
            <atom:updated>2020-06-07T17:14:53.497Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*U6N_ji0Ez8Ln0VzTUT2Mdg.png" /><figcaption>Credit: L.J. Stroemsdoerfer (no artist)</figcaption></figure><p>Machine learning deployment is definitely a hot topic right now. I was about to use the so-hot-right-now meme to back my first sentence, however, I feel everyone in the industry agrees with me on this. There are many cool approaches to deploy and operationalize ML models. However, they mostly introduce new tools, with their own ecosystem to work in. This adds a new layer of complexity and eventually might create lock-in effects, as you need to adjust your code to optimally fit the platform.</p><p>Open source concepts, such as Docker and Kubernetes grant much more freedom on this matter, however, are quite complex to use. Thus, I built a Python lib that allows you to professionalize, test and deploy your model on Kubernetes, however you don’t have to leave Python for it. Sounds too good to be true? Well, let’s see what you think.</p><h3>The productionize library</h3><p>Apart from the shitty title of the package — suggestions are highly welcome — the package allows you to do some pretty neat things to ease and speed up your deployment experience.</p><p>Let’s first take a look at what the library with the catchy name does.</p><h4>What does it do?</h4><p>Brief answer: it allows you to containerize your ML model without leaving Python. On top of that, it builds a local Kubernetes cluster on your machine to test if the containerized model is working properly. If you are thinking, wait, I don’t have a local Kubernetes cluster, don’t worry, I built in some functions to take care of this. Let’s take a closer look.</p><p>To begin your journey with productionize, you can first set up your local “workbench” as I called it. This is pretty much a Minikube cluster with a Docker registry running in a VM on your machine. To anyone who has set up such a stunt before, knows that this is a bit annoying to do. With productionize this is two lines of Python code.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FWWlzvIVJ_Y37HGO1QTVjA.png" /><figcaption>The workflow with productionize.</figcaption></figure><p>Once you have this “workbench” running, the productionize library expects a ML model that is already built into an API structure. You can use e.g. Flask for this, but also other tools. However, I would definitely recommend Flask, as I really like its lightweight approach.</p><p>With a few lines of Python code, you can then containerize your API and deploy it to the local workbench to test it. If everything is fine you can then push the container to any Kubernetes cluster you might like. So, let’s take a look at the code.</p><h4>How can you do that?</h4><p>First, you will need to install the library, which is hosted on PyPi. I am currently not supporting conda, please forgive me for that.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/93f958285635fcba546210767a2831e5/href">https://medium.com/media/93f958285635fcba546210767a2831e5/href</a></iframe><p>After that, you can import the library. Please be aware, that the library currently does not support Jupyter notebooks, as I am not using the Jupyter Kernel to setup the VM, the Minikube cluster and all the other nitty gritty components.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/ee2e0f5a94c9475d1fc0a7488ffdf4e1/href">https://medium.com/media/ee2e0f5a94c9475d1fc0a7488ffdf4e1/href</a></iframe><p>The library contains two major classes. The first one is the workbench() class, which takes care of the tools you need on your machine to do all those amazing things. The second one is the product() class, which handles your API and it’s containerization.</p><p>Let’s start with setting up the workbench. The workbench() class will install, manage and uninstall Docker, VirtualBox, Kubectl and Minikube for you. First, we initialize the class. This will check your local machine for the said components and mark those missing to be installed.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/040ed148b1546c431c19736ec513292b/href">https://medium.com/media/040ed148b1546c431c19736ec513292b/href</a></iframe><p>Now we can install the missing components and configure them to work well with each other. In the next steps, you will probably need to enter your sudo password.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/16f585ed9c5472f8be56140fcc113438/href">https://medium.com/media/16f585ed9c5472f8be56140fcc113438/href</a></iframe><p>If something goes wrong in this step, I added a debug method for this. Let’s say the installation of Minikube failed, you can just run:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/a5734c758e7ab2243691d8e98c0919fd/href">https://medium.com/media/a5734c758e7ab2243691d8e98c0919fd/href</a></iframe><p>Once the setup step is completed, you can fire up the workbench, which essentially fires up the Minikube cluster on your machine.</p><p>Next, you should create a project on the workbench, which creates a Kubernetes namespace. I would advise you to do so, this creates some sense of order on your cluster.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/9c244cdc5474c9a3d982b640d222577b/href">https://medium.com/media/9c244cdc5474c9a3d982b640d222577b/href</a></iframe><p>If you are sick of your workbench you can easily stop the cluster using the stop_cluster()method and you can even cleanly remove all the unwanted components using the uninstall()method.</p><p>Let’s now move on to the product() class. This class turns you ML API into a container and deploys it to the workbench. First, we have to initialize the class:</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/cea9554a8712d97ec91000bba0136796/href">https://medium.com/media/cea9554a8712d97ec91000bba0136796/href</a></iframe><p>You can now prepare the deployment, which will essentially create a Dockerfile. I added this intermediary step so that you can edit the Dockerfile, if you need.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/3da989743def2077ecca1a575c03a0e2/href">https://medium.com/media/3da989743def2077ecca1a575c03a0e2/href</a></iframe><p>Whether or not you edited the Dockerfile, you can now deploy it to the local workbench using the deploy() method.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/581051e481b17cd455688fd43f3773fc/href">https://medium.com/media/581051e481b17cd455688fd43f3773fc/href</a></iframe><p>This will create the Docker image on the Minikube registry, deploy a pod with your container and expose the service to external. In the output of the deploy() method, you will find the link to your API. You can use this to test, whether your model works just fine. If it does, you can push it to any Kubernetes cluster you want. You can do so using the push_product() method and pass the registry link of your Kubernetes cluster.</p><iframe src="" width="0" height="0" frameborder="0" scrolling="no"><a href="https://medium.com/media/5decb87c1a7831a89ec1708d9094e3ca/href">https://medium.com/media/5decb87c1a7831a89ec1708d9094e3ca/href</a></iframe><p>Okay, so that was a lot. I know this is a bit of a complicated structure, so I summarized it again.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*CCFt9DhO8vydzDSMWY1oRw.png" /><figcaption>productionize workflow at its best.</figcaption></figure><p>There are additional information and documentation on GitHub and PyPi. You can check it out. Also, there are of course some funny easter eggs included.</p><h4>What is next?</h4><p>As you see, the library provides some nice features, however, it is far from being completed. At the moment, productionize only works on macOS. This is due to the fact, that I am using Homebrew as a main package manager. I am already working on a version for Ubuntu. Any linux distribution will be manageable, however, Windows is going to be quite a challenge. So if you have suggestion or ideas, please let me know. Also, I would be more than happy to work on any bugs, if you drop me an issue on <a href="https://github.com/LJStroemsdoerfer/productionize">https://github.com/LJStroemsdoerfer/productionize</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e64a37c3155a" width="1" height="1" alt=""><hr><p><a href="https://medium.com/data-science/machine-learning-deployment-with-python-e64a37c3155a">Machine Learning deployment with Python</a> was originally published in <a href="https://medium.com/data-science">TDS Archive</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>