Introduction

Artificial Intelligence is reshaping the way we work, connect, and innovate.

At the heart of this transformation is Research and Development (R&D), where AI is not just a topic of exploration, but a practical tool driving innovation. In today’s article, we spotlight Alicja Kasicka, an R&D Specialist at Orange Innovation Poland, whose journey exemplifies how AI, research, and creativity converge.

Alicja studied Artificial Intelligence at Poznan University of Technology and spent a year in Luxembourg, gaining international experience and exposure to AI applications.

Her early curiosity about critical thinking and research deepened throughout her studies, where she discovered that data is more than just numbers—it's a tool to understand, create, and question. Currently, she works in the AI competence center at Orange.

The Career Path: From Curiosity to AI Enthusiast

Alicja’s road to AI began in high school, where she focused on biology, chemistry, English, and mathematics. While considering her next steps, she discovered the competitive AI program at Poznań University of Technology and decided to take on the challenge.

It turned out to be the right fit.

"Studying at Poznań University of Technology helped me discover my passion for data. Choosing AI turned out to be a great way to combine my analytical mindset with something practical and impactful."

Then, her journey into research started with the Hi-Tech Girls program at Orange, where she applied through the Data & AI track. As part of the recruitment process, she had to design a recommender system for Orange. That challenge led to an internship as a Data Scientist, and eventually to a permanent role in R&D.

AI in Action: Projects and Lessons from Industry

During her time at Orange Innovation Poland, Alicja contributed to several NLP projects that showcased the practical challenges of applying AI in business contexts. One of her projects focuses on evaluation of chatbot’s answers in human-chatbot dialogs.

"Quantitative evaluation is pretty straightforward, but qualitative one is tricky. For example, how do you measure empathy in a chatbot's response?

That’s the kind of challenge we are trying to tackle."

These experiences reinforced the importance of understanding both technical implementation and human-centered design.

For Alicja, it’s not just about the technology- it’s about making sure created solutions work in the real world, with reliable data and a clear understanding of human behaviour.

Studying Abroad and Expanding Perspectives

Spending a year in Luxembourg added a global dimension to Alicja’s journey. She noted how the culture there encourages equal opportunity - even Erasmus students can conduct their  own research projects.

"I had a chance to take part in an AI & Art project, but at the time I was already working at Orange, so I couldn’t commit. Still, the experience taught me how important general knowledge is when working in AI."

Her time abroad broadened her understanding of AI’s role across industries, reinforcing the importance of interdisciplinary knowledge and global collaboration in research.

Key Trends in AI - What Alicja thinks is crucial?

Looking to the future, Alicja sees several technologies gaining momentum and reshaping how AI is applied across industries:

• Intelligent Agents - AI-driven assistants capable of increasingly autonomous behavior in customer service, enterprise systems, and beyond.

• Extended Reality (XR) and Virtual Reality (VR) - These tools will go beyond entertainment, supporting onboarding processes, immersive learning, and accessibility -for example, helping people with disabilities explore museums or participate in virtual classrooms.

• Digital Twins - Digital representations of physical environments like factories or cities, useful for remote training, testing, and smart city infrastructure.

"XR is not just a buzzword. It has real potential in smart cities, education, and accessibility. The degree of implementation will highly depend on the industry."

These developments reflect a broader shift toward making AI more embedded, interactive, and impactful in real-world environments.

Ethics and Responsibility in AI

Ethical AI is one of the central concerns in Alicja’s work. She emphasizes three pillars: transparency, responsibility, and most importantly, context awareness.

Understanding the intended use case is essential - not all models are equally appropriate for all scenarios. A model with great performance in a controlled academic benchmark might fail dramatically in real-world, high-stakes environments.

Alicja also highlights the importance of checking how performance metrics are distributed across all classes and how misleading aggregated statistics can be.

"I once found a model that claimed 90% accuracy. Sounds great, right? But it turned out that this 90% accuracy came from just one class - meaning it worked like a binary classifier instead of recognizing five categories."

Common Mistakes and Challenges in AI

According to Alicja, beginners often fall into similar traps:

1. Overlooking data quality: Clean, relevant data is foundational. If your dataset is flawed, your model's predictions will be too.

2. Lack of high-quality data: Models can appear to perform well but may suffer from overfitting or hidden bias. A classic example is a wolf vs. dog classifier that just learned to recognize snow instead of the animal.

3. Using complex models without understanding them: Jumping straight into advanced models without understanding how they work can lead to misinterpretation. For example, forgetting to adjust the embedding tokenizer space from English to Polish can make your model completely ineffective.

4. Focusing on hype: Chasing after the most “popular” tools or models without theoretical grounding can limit growth. Theory matters—without it, you can't build or adapt models effectively.

She also highlights challenges specific to the research and business landscape

• Keeping up with global AI trends - The field moves rapidly, and it's impossible to follow every trend. Her advice? Focus on your specific domain and don’t try to chase everything.

• Unclear research goals - Many researchers work with a large number of parameters but lose sight of their hypothesis. It's important to limit your research scope and validate specific assumptions.

Advice for Students and Junior Researchers

For those entering the world of AI and Research & Development, Alicja shares valuable insights:

“Coding skills are still relevant - but equally important are critical thinking, verifying sources, and asking the right questions.”

She emphasizes that practical skills and mindset go hand in hand. Theoretical knowledge is vital to truly understand when and how to use a specific tool or model.

She also recommends:

1. Learning prompt engineering, which allows for more effective interactions with large language models and makes it easier to prototype ideas or test assumptions

2. Developing soft skills, especially storytelling and communication, which are crucial in both research and industry settings

3. Building broad general knowledge, since many challenges in AI require connecting dots across disciplines

4. Avoiding overreliance on trendy tools without understanding the underlying theoretical foundations

Alicja’s story reflects the evolving role of AI in both industry and research.

With a strong foundation in Research & Development and a passion for making technology, she is helping shape a future where AI is used responsibly and creatively.