Data Scientist | Machine Learning Engineer | Data Engineer

6/2024 – 7/2024

Delivered an emotion analysis of reviews for their client - an internationally known group within beauty and personal care. By performing this analysis it
was possible to identify growth opportunities for certain categories, products and brands by looking at, e.g., emotion trends over time. The delivery consisted of data analysis conclusions, visualizations, and workshop material for the end-client to use. Web scraped reviews were emotion-classified by using the RoBERTa Emotion Base model (utilizing the same underlying architecture as in many generative AI models) and then the emotion scores were processed using common data processing tools. This has provided the client with a type of analysis that previously was not feasible, and it provides a new dimension to their customer insights.

Data Science, Python, Microsoft Excel

4/2021 – 4/2023

▪ Engineered end-to-end AI/ML solutions, overhauling existing frameworks to provide the assortment team with actionable data for optimized
decision-making. This involved assortment optimization, vendor negotiations, and tracking campaign success, contributing to annual cost savings exceeding 100 MSEK.
▪ Developed high-precision classification models (AUROC > 0.95) for professional customer types, reshaping targeted promotions and assortment optimization for my counterpart.
▪ Refactored and improved a deep neural network regression model used to enhance the measurement of promotion effects by assessing variances in non-promotional sales.
▪ Orchestrated workflows and ETL pipelines spearheading a seamless system migration to ensure reliable and timely data access. This included advanced stored procedures written in SQL, as well as defining tables, views, and triggers.
▪ Campaign data analysis to determine if they had the desired effect. This included to take, e.g., baseline sales and cannibalization, into consideration.
▪ Created consumer decision trees as a data driven way of visualizing the customer purchase process in a specific product category which supports the category manager in deciding on the assortment.
▪ Created customized reports, metrics and data visualizations as a part of the data analysis and communication process.
▪ Mentored a junior data scientist, actively contributing to team growth and knowledge transfer within the organization.

Amazon Web Services (AWS), Apache Spark, Subversion, Git, Jenkins, Power Bi, Microstrategy, Python, R (Programmiersprache), Scikit-learn, SQL, Tensorflow, Teradata Sql

3/2019 – 4/2021

▪ Experimented, developed, productionalized and performed operations and maintenance of AI/ML models for large scale transaction monitoring at the Swedbank Group and thus, contributing significantly to the advancement of its AML/CTF capabilities. This was performed on the data lake by utilizing Apache Spark and Apache Hive (SQL).
▪ Reviewed a use-case where semi-supervised GANs were used to generate synthetic ML/TF patterns that could be included in the transaction monitoring process. GANs are a powerful and flexible tool in the field of generative AI, capable of creating highly realistic synthetic data across various domains
▪ Drove the implementation of the 1st unsupervised anomaly detection model, integrating Know-Your-Customer data.
▪ Guided a junior data scientist and co-developed an anomaly detection model targeting transaction behavior in a popular instant money transfer service.
The new solution led to substantial improvements compared to an old rule based solution.
▪ Created customized data visualizations using common tools like Shiny, Plotly, Dash, Matplotlib, and Seaborn to support the data analysis and
communication process.
▪ Co-authored and published a paper, actively participating in the development, testing, and real-world application of a graph-based solution for detecting
anomalous group behavior, addressing challenges with incomplete network information.

Apache Spark, Git, Jenkins, Microsoft Azure, Python, R (Programmiersprache), Scikit-learn, SQL, Tensorflow, Teradata Sql