Meet Gabriel:

Keeping pace with a rapidly changing industry

This man of many interests, highly driven and intrinsically motivated, has found his place in the world back home, in the city of Bucharest and at Endava, as Principal Data Scientist.  

“I was doing research and founded a company to do applied research and contract work while working for a smaller company before joining Endava. So, I did all three at the same time,” explained Gabriel.  

He admitted this was one of the two start-ups he founded and ran to the ground because other endeavours were also competing for his attention. However, he does not seem to mind. “It’s ok,” he said, appreciating the gained experience. 

As someone who loves to explore different areas, learn and build what had not been built before, Gabriel is always exploring new horizons. Having trained as an Electrical Engineering Specialist, he joined Endava more than 10 years ago as a Project Manager before later becoming Senior Project Manager. His next move would define his career to this day.   

“In 2015, I started to move my research interest to data science,” he said. “In 2017, I started to become more involved in data science activities in Endava, organising the data science community and delivering training in data science.” 

This was the beginning of a new chapter in Gabriel’s story that reflected larger changes across the industry. Data had long been a commodity, but its worth was relative to the emerging science behind it. At this point, its capacity for use in seeing and predicting patterns was increasing in speed and accuracy.   

“During my tenure as a Data Scientist with Endava,” Gabriel continued, “there were many breakthroughs in the fields of data science and machine learning (ML). One important milestone was in 2017, with the publishing of the seminal paper from Google on transformers (the basic architecture for today’s large language models), called ‘Attention is All You Need’. This led to a revolution in the application of natural language processing for various tasks.” 

Other milestones include various AI releases and technology advancements. Being at the forefront of any field in IT – especially one as fast evolving as AI, ML and data – requires commitment, and it’s something Gabriel prides himself on.  

“I follow closely the latest developments in generative AI,” he said. “I am still very much involved in competitive predictive modelling, and I constantly publish content – mainly notebooks on Kaggle. I read about several domains related to data science, ML, AI, including MLOps, ML systems design, natural language processing (NLP), but also computer vision and recommender systems.” 

In passing, Gabriel modestly added: “I also published a book last year on advanced data analytics, predictive modelling and generative AI.”  

 
His book, Developing Kaggle Notebooks: Pave your way to becoming a Kaggle Notebooks Grandmaster, is available now on Amazon. 

Teamwork and community spirit

For Gabriel, it’s important to always seize opportunities to learn. At Endava, he’s found a place where he can do just this by creating communities. Along with the Agile Community, he has also developed most of the initial materials for the Risk Management Community and organised the AI communities.  

“At first, they were people from all different backgrounds and skills, getting together, talking. Engineers need this kind of environment where they can discover new ideas, places where they can discuss new trends. Of course, they need to learn a lot by doing the project work. But we can do it like at Endava. 

“It's important not only to learn by yourself, but to be part of a community of learners, right?  One thing that I brought with me was that I knew how important communities are.” 

Another opportunity Gabriel likes to seize at Endava is the Innovation Lab – and he’s done so to great success. 

“I’ve taken part twice in Innovation Lab. A few years back, in 2018, when my team got third place in the finals, and last year (2023), when I also reached third place. Both times, we started with a cool idea of applying ML to a real-case scenario. 

“In the first project, we built several ML models and trained them with internal requests and incident data to predict various ticket attributes. The solution we presented in the finals was integrated with our internal information technology services. After calling the endpoints exposing the models, the component we integrated from Endava’s automation platform could prefill several fields in the user interface (UI) displaying the current ticket. The human operator only needs to validate the automatic classification and continue to work on the ticket. 

“For the second project, we started with the idea of developing a solution to make internal information in Endava more accessible through the use of generative AI. Our approach used a retrieval augmented generation (RAG) system. From the local phase to the finals, we evolved our initial solution. 

“After discussions with two industry verticals, we developed two different products: [the first was] a medical insights chatbot, that will answer questions about diseases and symptoms. For this, we integrated the solution with a medical terminology thesaurus to provide data enrichment to the answer returned by the large language model (LLM) to the user queries. For the banking and capital markets, we created a complex agent that manages multiple specialised agents that facilitates their cooperation for addressing complex tasks.” 

Outside his life at Endava, Gabriel shares his knowledge and stories in books and articles, through communities and presentations.