Follow Us
Machine learning engineering & data science team is the team behind Velti’s machine learning product and AI/optimization initiatives.
Our team is responsible for the product development of Daedalus, Velti’s AI based engine, which uses machine learning algorithms for next-best-action prediction and optimization of various targets.
From requirements analysis to design and implementation, our team handles all the steps of building a product that helps us operationalize and use ML models to make decisions.
We translate business problems to ML pipelines and optimization – with Daedalus product as a basis, we also handle the launch and continuous optimization of ML projects for major mobile Operators in Europe and worldwide.
Machine learning engineering & data science team is the team behind Velti’s machine learning product and AI/optimization initiatives.
Our team is responsible for the product development of Daedalus, Velti’s AI based engine, which uses machine learning algorithms for next-best-action prediction and optimization of various targets.
From requirements analysis to design and implementation, our team handles all the steps of building a product that helps us operationalize and use ML models to make decisions.
We translate business problems to ML pipelines and optimization – with Daedalus product as a basis, we also handle the launch and continuous optimization of ML projects for major mobile Operators in Europe and worldwide.
For each project, our team builds an automated, self-learning system that involves data ingestion, feature engineering, provision of recommendations, retrain of the ML models.
Finally, we invest time in the various steps of the modelling process: evaluation setup and sampling/weighting techniques, experimentation with different algorithms and stacking, hyper-parameters’ tuning and model evaluation.
Applied machine learning is always a challenge: our results are measured against specific KPIs or competing with traditional strategies of business experts.
The power of our team comes from combining diverse backgrounds, including computer science and engineering, data science and mathematics.
We work with Java and modern frameworks like Spring Boot, in a Microservices architecture that allows us a productized approach: reusable, stable components, easily extensible in a plugin architecture.
We utilize Python, machine learning libraries and platforms to train our models. A variety of different tools like H2O AI, Airflow and Apache Spark gives us flexibility and performance.
We experiment with different algorithms, NLP techniques and text optimization. And finally, we use tools for model interpretability, which allow us to understand our models, provide feedback to the business teams, and adapt optimization strategies.
Our team takes pride in the code and the machine learning models that we build, puts emphasis on collaboration and teamwork spirit, and has a customer-driven approach.
We collaborate with different departments, within Velti and with clients. We work closely with the business, sales teams and BI to design and continuously tune the optimization running for each project. Our team’s work contributes to expanding the use of machine learning throughout Velti’s projects and empowering data-driven decisions.