In an interview

Andreas Luckert - AI Engineer

Andreas Luckert - AI Engineer

About Your Job

What does your typical workday look like?

Since I’m definitely not a morning person, I start work no earlier than 10:00 a.m., either from my home office or at 10:30 or 11:00 a.m. at our beautiful office in Heidelberg. I don’t mind staying late in the evening - it just fits better with my personal life. I’m actually always at the office because I really appreciate being close to my colleagues and the atmosphere there. Because of early morning meetings, I’m sometimes at home before lunch, but then I head to the office.

What makes your team special?

The team is spread out internationally across several continents - at least on the client side. There are daily stand-ups, and in my role as Technical Lead for one of the teams on this large-scale project, there are additional scheduled meetings. Regular contact with the developers on my team and the other teams in the project is essential. I prefer asynchronous chatting in MS Teams rather than too many ad-hoc calls or “alignment meetings,” since otherwise you’d hardly get any work done yourself and would be pulled out of your flow too often. Exceptions are “more personal” topics - in those cases, a call is the method of choice.

What makes working as an AI engineer so exciting?

I enjoy automating more and more of the “boring” tasks and fine-tuning the “agent harness” until you hardly have to say much in the prompt anymore, because the skills are all formulated so clearly and precisely that the AI agents act on your behalf in a fairly “deterministic” and controlled manner, allowing you to focus on more important, more enjoyable, and more interesting things - things that have more to do with creativity, software, product, and architectural design, as well as interpersonal communication.

What development in the field of AI has surprised you the most in recent years?

The breakneck speed at which these tools have been integrated - from simple “browser-based chatbots” to “agent-like features” embedded in everyday applications. Of course, the big players (OpenAI, Anthropic, etc.) have all been living off VC money and have therefore kept token prices artificially low for a very long time so that we could all learn what to do with them - but also become dependent on them. Now, all of a sudden, we have to be more frugal and watch how much we use and for what. Once you get the hang of it, I think it’s generally a positive development because, in the medium and long term, many tedious tasks and activities will no longer have to be done by humans. From my own perspective, I also understand why this can sometimes be a cause for concern, but overall, I’m actually grateful that I can now delegate so many things. I’m very curious to see how this will play out in the coming years.

Which project do you remember most fondly and why?

For many years, I was primarily involved in a major project that I look back on fondly and that I’m still happy to be a part of, because it’s far from over! What I especially like about the project is the collegial, positive, and international team atmosphere. I also appreciate that you’re always learning something new, and that the project consistently stays on top of the latest technologies and possibilities. Thanks to my many years of involvement in the project, I’ve had the opportunity to play a major role in shaping it - as well as its culture, work processes, and the product itself.

What have you learned through your work on AI that many people wouldn't expect?

When it comes to this question, I probably need to put myself in the shoes of people who don’t live and work in the same bubble as I do: What I hear time and again is that “the models aren’t that good” or “they also spout a lot of nonsense.”

These days, especially given the level of today’s LLMs, the responsibility in most cases no longer lies solely with the model, but with the person who provided the prompt.

In addition to pure prompt engineering (a term that’s been around for about three years now), there’s also what’s known as “harness engineering,” in which the agent or LLM is given a clear, predefined framework within which it is expected to act in specific ways in familiar situations or when performing known tasks. This way, you can avoid those “annoying mistakes” - such as “I’ve already said this a million times; it should do it this way or that way” - by distilling these learnings into “Agent Instructions” and “Skill Files,” for example, and documenting them in a concise, AI-ready format.

By proceeding iteratively in this way, you can eventually rely more and more on the AI agents to perform known tasks correctly and immediately within the desired framework, without having to make repeated corrections.

Working in an Area of Expertise

What does working in an area of expertise mean to you?

It’s a great opportunity to explore new and relevant technologies, stay up to date, and keep learning as the field evolves especially if that isn’t possible to the same extent in client projects.
If you’re lucky, this goes hand in hand with day-to-day business on client projects, since clients expect you to constantly engage with technological innovations, especially when they’re relevant to the project.
In addition, through working in an area of expertise, I’ve gotten to know many other HMS employees with whom I otherwise would have had little interaction. So, in addition to the primary goals of building and exchanging knowledge, it also serves as a team-building activity.

You're part of the Cloud Engineering area of expertise. How do you contribute to this?

I really enjoy working with our active team members to shape the direction and practical implementation of our activities in this area of expertise - improving and expanding our existing internal AWS applications, launching new pilot projects, and inspiring others to engage with AWS-related topics.
Once or twice a year, I also get to travel to Hamburg, Berlin, or Munich for one of the major AWS events, which is, of course, really cool.

What are the most important topics you're currently working on?

“All things AI” may sound a bit overused by now, but it’s simply the top topic in the cloud as well: The already well-established Amazon Bedrock service, which provides everything related to LLMs, has been complemented for just under a year by the AgentCore service, which, as the name suggests, is about “Agents as a Service” (or a platform).

A month ago, as part of a project in the area of expertise, we implemented proof-of-concepts on AWS in which we developed agent-based apps for HMS employees from all departments that simplify everyone’s daily tasks and help everyone stay up to date.

In addition, over the past three months, I’ve developed an “Agentic Software Development Lifecycle” portal, or ADLC portal for short, which essentially abstracts the entire technical process away from the end user. All the user has to do is enter a detailed description of their desired app into the portal, and then everything is implemented, tested, deployed, and made available in a front end - all “as if by magic.”

We’re currently testing this internally across several departments to see if it’s a viable way to collaboratively further develop our company applications in the future and also to design entirely new apps that otherwise would never have become a reality.

What are your responsibilities in the Cloud Engineering area of expertise?

Since the beginning of this year, I have been the head of the "AWS division" within the new Cloud Engineering area of expertise and the AWS Product Owner at HMS. 

What developments do you foresee in your area of expertise over the next few years?

I think that agent- and LLM-based services at AWS will continue to grow in importance. Nevertheless, we must not neglect the core services that hold everything together at AWS and in the cloud in general. That’s why solid cloud training through our internal training programs and official certifications will always remain relevant, because for many years to come, it will still be necessary to have a good understanding of why a particular architecture is the right one, in which cases which services should be used, and so on. As things stand, this cannot simply be left to agents to handle blindly, and in my opinion, that won’t be the case for the foreseeable future either. 

Staying on top of things, continuing to learn, and putting new ideas into practice is just plain fun.

Andreas Luckert

AI Engineer

Workation and Remote Work

You've already been on a few workations. What motivated you to do that?

I’ve always had a healthy dose of wanderlust in my blood, and when I realized in my mid-20s just how big and diverse the world really is, I did everything I could to spend as much time abroad as possible while I was still in school - especially because I believed that as soon as I started working, my travel would be limited to 30 days of vacation per year. Back then, I hadn’t even considered working in the IT industry as a software and cloud engineer. Fortunately, that’s how things turned out, and I was even luckier, of course, with HMS, which allowed me - right after my probationary period in 2021 - to work from Chile for about three months, spanning the winter of 21/22. I personally love living and working here in Heidelberg, but especially when winter comes, my wanderlust gets the best of me, and with this combination of variety and a warm climate in winter, I just really love getting away. Of course, there’s also a bit of an adventurous element to it, and you can discover more of the world than you can with “just” the standard vacation time you get each year. The fact that this was even possible while doing my job was, of course, fantastic, so I was grateful to be able to push this initiative forward with a lot of support at HMS. We now have a well-established process, and many employees are taking advantage of it.

How do you manage to balance work with experiencing new things while you're away?

You have to admit quite clearly that it’s especially on weekends and holidays or even during short breaks that you’re most able to explore new and unfamiliar places and enjoy them to the fullest. After all, even though “remote work” is often marketed as a “workation” as if you were on “vacation”, people forget that you’re still (in my case) working 40 hours a week, with the corresponding meetings and deadlines on your calendar, and so on.

But even during the week, if you’re creative with your flexible work schedule, you can still get a lot out of the place: For example, when I was in Alicante, my mom (who, for once, was there for a month) and I would always explore the city and the surrounding area between 2 and 6 p.m. and soak up plenty of sun and fresh air. When the sun went down at 6 p.m., we’d head home, and I’d keep working for a good three hours. Of course, this isn’t for everyone, but it worked great for me because it allowed me to see a lot of the area in daylight and not be limited to just the weekends.

On the other hand, if you stay in the same place for about five months, like I did, it’s kind of like a “Work-Erasmus,” where you can develop hobbies, routines, and friendships. That’s obviously a completely different experience—less touristy and more about routine and genuine connections with other people. Another plus: you’re not under as much pressure to do and see so much every day, because you have plenty of time.

How does effective collaboration work in a team that works remotely?

Incidentally, in my line of work, this isn’t something specific to “remote work” - it’s more the norm: Even when I’m at the Heidelberg office every day, all my project team members are always scattered around the world. At least in my situation, it only happens occasionally that other HMS colleagues working on the same major project are also in the office. It’s more common with projects within HMS to have on-site meetings. This approach works particularly well in such distributed teams when you strike the right balance between in-person interaction and discussion in targeted meetings, as well as effective asynchronous communication via Teams channels and chats.
 

What do you appreciate most about working remotely?

Especially in winter, I like to visit sunny places to experience more variety and a sense of adventure, and to create memories that stand out more clearly. For me, a change of scenery is simply inspiring and invigorating. New surroundings and experiences always lead me to a completely different kind of self-reflection, which gives rise to entirely new ideas. When I return to Heidelberg, I appreciate everything that’s so great about this place much more than if I were here all the time.

Ever since I read that the human brain is an efficiency machine that heavily condenses or compresses all similar memories and experiences, and, in turn, stores all very different experiences clearly separately, thereby creating more distinguishable memories, I’ve had one more reason to make remote work a regular part of my routine. I just don’t want to have to say in 10 or 20 years: “The years have flown by so quickly, and I can barely tell them apart in my memory.” Changes in surroundings and culture are definitely different enough that the brain can’t simply compress these experiences. So, looking back, it feels as though you’ve experienced so much more. A true life hack!

What challenges does working from abroad present? And how do you overcome them? Do you have any specific tips?

Setting up your workspace is the biggest challenge during work hours: If it’s not ergonomic, eight hours can feel like an eternity, especially over the course of several weeks and months. Other challenges mainly involve organization, paperwork, accommodations, flights, subletting your apartment, pausing all kinds of subscriptions and memberships, having your mail forwarded via a “mail forwarding request,” and so on. Dealing with the local language, only being able to see a doctor in an emergency for several months, finding new hobbies locally, figuring out where to go for what, registering and deregistering for various services, scheduling all doctor’s appointments to coincide with your time in Germany, and much more. Yes, that’s a lot of overhead and not for everyone. But I’d do it all over again. And you get faster and better at everything; it’s all a matter of experience and, after several years, it becomes a kind of routine.

Why do you think workation is an attractive benefit?

Phew, I think that’s clear from the points above.
From an employer’s perspective, I’d like to add that every company would be well advised to offer something like this, because it tends to boost employee performance as well as employee loyalty to the company and satisfaction both at work and in life. In my opinion, it doesn’t make sense to force people to be in a specific location all year round for jobs where it isn’t absolutely necessary. I’m glad that HMS was open to this idea five years ago, and that it has now become standard practice at our company.

Rhein-Neckar AI & Data Meetup

You're co-organizing the Rhein-Neckar AI & Data Hub Meetup. How did the idea for it come about?

It all started at the AWS DACH Community Event in mid-September 2024: There were posters everywhere promoting the so-called “AWS User Groups.” There were already some in Karlsruhe, Heilbronn, Frankfurt, and Bensheim, but none in Heidelberg or Mannheim. So, just one week after the event, I went to a user group meetup organized by the Heilbronn group. I enjoyed it so much that I thought, “Wow, we have to organize something like this in Heidelberg and the surrounding area, too!”

The idea of starting an “AWS User Group” eventually evolved into the much broader and more independent “Rhein-Neckar AI & Data Hub Meetup.” The very first meetup took place on February 20, 2025, while I was working remotely from Málaga; so I couldn’t even attend the first event in person, even though I’d played a major role in organizing it. By the second event in April 2025, I was back in Heidelberg and immediately took on the role of host. This has since become the norm. I always take on the role of the host, unless, in rare cases, I’m unable to be there. It’s incredibly fun, and by now, after a total of nine meetups, my stage fright has really subsided. 😉

So my responsibilities mainly involve hosting the meetups, reaching out to external and internal speakers, and brainstorming with the team about how we want to organize future meetups, including topics, supporting programs, and formats, etc. Thanks to my visibility on LinkedIn, I’m also regularly contacted by potential speakers as well as applicants for HMS. So I also handle this external outreach and the management of these channels on the side.

What is the purpose of the meetup?

To create a space in the Rhine-Neckar region where professional expertise and “curious minds” can come together to exchange ideas, learn from one another, and inspire each other. Of course, increased visibility for HMS both in the region and in the real world, as well as within professional networks, particularly on LinkedIn, is a positive side effect of all this. After all, ultimately, several HMS employees have to invest their time and energy into the whole endeavor to keep it running. That’s why we’re also regularly looking for other like-minded individuals who’d like to help us organize these kinds of meetups. In the past, for example, we’ve organized events with SRH University.

Why is networking and collaboration within the data and AI community important to you?

While you can learn a lot online through tutorials and blog posts, ultimately, there’s simply no substitute for a live presentation followed by a discussion and an informal exchange of expertise. The connections, opportunities, and insights you gain from such events should not be underestimated.

What can visitors expect from the meetup?

Highly informative presentations, lively and respectful discussions and Q&A sessions, as well as casual networking over delicious food in a beautiful setting overlooking the rooftops of Heidelberg’s Bahnstadt district.

What kind of experience or skills should visitors already have?

In principle, anyone interested in a topic is welcome to attend, regardless of prior experience or skill level. However, some of the presentations - and especially some of the questions during the discussion - can sometimes get quite technical and specialized, so it’s a good idea to have at least a basic understanding of the topic. 😉 The meetups are always held in English because we always have a lot of international visitors, and we’d hate to exclude them. During the networking session afterward, German is also spoken, depending on who you’re with in your small group, as long as everyone’s okay with it.

How exactly does the meetup work?

We’ll start with a welcome session and some casual mingling. Then, after 30 minutes, the first talk will begin, followed by a Q&A, and then the second talk, also followed by a Q&A. Finally, snacks will be served, kicking off some casual networking.

What did you personally learn as a co-organizer of the meetup?

From pure inspiration to an idea, then to a vision, and finally to concrete implementation. Looking back, that’s actually the most amazing thing about it all - that it really worked out and was so well received within the company. Because without that support, it wouldn’t have worked. It also gave me insights into event management, the challenges that come with it, and how to overcome them. Hosting the meetups has certainly been a standout experience and personally very enriching for me, since you rarely get opportunities like that in day-to-day business.

About HMS

Why did you choose HMS as an employer back then?

At the time, I found the website significantly more appealing, modern, user-friendly, and inviting than almost any other I had seen. The fact that you could even submit a proactive application and that there was a dedicated subpage for it was the icing on the cake. The relaxed and friendly conversation with Kai Brodmann, as well as the prospects of everything I could learn here about cloud and software engineering, were reason enough for me to take the plunge and make the career switch from physics to IT consulting.

To this day, I haven’t regretted it; there are plenty of interesting projects and learning opportunities, and from a technological standpoint, you’re always at the cutting edge. As an IT consulting firm, we’re, by definition, always required to stay several steps ahead of our clients, and this constant need to stay on top of things, keep learning, and implement new ideas is just a lot of fun.

Why are you staying at HMS?

On top of that, there's the great, friendly, and relaxed atmosphere among coworkers in the office, as well as the opportunity to work with a high degree of flexibility in terms of both time and location.

In your opinion, what sets HMS apart from other employers?

The fact that at HMS, hierarchies are flat and communication lines are short, and there’s a strong spirit of innovation and flexibility. Ideas like the meetup or workation would most likely have been nipped in the bud at large corporations, where everything is regulated down to the last detail and you interact with other departments - if at all - only through an anonymous ticketing system. You have to be aware of this when deciding whether to work at a large corporation, a medium-sized company, or a startup. They all have their pros and cons, and I clearly see the advantages here and knew how to leverage them for myself, the company, and my colleagues.