Work Experience
Principal AI Developer, Karnov Group AB
October 2024 - Present • Copenhagen, Denmark
As the Technical AI Lead for the Nordics, I led a team of 5 AI Scientists, working on KAILA - a GenAI agentic-powered legal assistant that I created and scaled to manage 3.5M documents. KAILA now has a team of over 50 people, consistently generates millions in revenue, and has been featured in multiple newspapers. It is used by top firms in both Denmark and Sweden.
Spearheading AI research and innovation, I collaborated closely with business stakeholders to ensure new initiatives were brought to market while ensuring research and implementation of state-of-the-art AI.
Tools & Techniques: PydanticAI, Langgraph, Python, Google Gemini, OpenAI, LocalLLMs, Docker, Google Cloud Platform (GCP), Generative AI, Information Retrieval, DeepSearch.
Senior AI Developer, Karnov Group AB
October 2022 - December 2024 • Copenhagen, Denmark
Created KAILA, an agentic GenAI-based conversational legal assistant, growing its document capacity from 5k to 3.5M and delivered in six months.
Grew the AI team from one to a unit of 5 AI developers and collaborated with several software engineers and business stakeholders across multiple teams. Overviewed the embeddings creation pipeline, database storage, and cloud infrastructure that supports KAILA.
Created several other PoCs, conducted several sessions to ensure knowledge sharing, and analyzed the efficiency of several million search queries.
Tools & Techniques: Transformers, RAG, Semantic Search, PyTorch, Information Retrieval, Kubernetes, Docker, Python, SQL, Machine Learning, Natural Language Processing (NLP), LangChain.
Senior Data Scientist, GoSimplo
November 2021 - September 2022 • Copenhagen, Denmark
In my role, I leveraged deep learning to build a prediction model using Keras and TensorFlow, which determined the probability of a project’s financial success through classification. Additionally, I conducted extensive research and developed several time-series and regression models to forecast various financial metrics, such as revenue and staff costs, utilizing XGBoost and ARIMA. These models were deployed using Docker via Azure Cloud.
I also created a robust training and deployment infrastructure for deep learning models using Python and Docker, and monitored their performance with Azure App Insights. My responsibilities extended to processing large-scale data using Azure Synapse and SQL, and providing actionable insights to stakeholders through dashboards built in Amplitude. Furthermore, I developed custom dashboards using data from Amplitude and Intercom, employing Pandas, Matplotlib, and Streamlit for visualization.
Tools: Microsoft Azure, Azure Synapse, Data Visualization, Financial Forecasting, Docker, Python, SQL, Machine Learning, Time series, Deep Learning.
Data Scientist, JohnConnor.ai ApS
June 2020 - November 2021 • Copenhagen, Denmark
JohnConnor.ai ApS was an AI/ML company I created with the purpose of providing services in the same area. In my role, I enhanced platform recommendations for GrantCompass by aggregating and visualizing data, and building a more effective recommendation engine using Python, Pandas, Spacy, AWS S3, and AWS Personalize. Additionally, I presented potential data science and AI-based options for GrantCompass by aggregating data from multiple sources and consulting with different departments.
For Karnov Group Denmark, I provided valuable insights through statistical analysis of user activity, utilizing Python, Pandas, Matplotlib/Seaborn, Jupyter Notebooks, and SQL, and effectively communicated these findings to management and marketing stakeholders. In the realm of data engineering, I improved data accessibility by aggregating multiple data sources, including APIs, MongoDB, and Oracle DB, and creating a data warehouse using Python, PostgreSQL, and JavaScript.
Tools: Amazon Web Services (AWS), Docker, Python, SQL, Machine Learning, Natural Language Processing (NLP).
Data Scientist, Karnov Group Denmark
July 2019 - May 2020 • Copenhagen, Denmark
I significantly enhanced recommendations for Danish-language documents by collaborating on content-based recommender systems. Utilizing Python, spaCy, Prodigy, Jupyter Notebooks, PostgreSQL, Docker, Kafka, and Kubernetes, I successfully improved the relevance and accuracy of these recommendations.
Additionally, I worked on a PoC with the goal to optimize the platform’s search functionality by annotating data and developing a Deep Learning model focused on law-specific terms, employing Python, Prodigy, spaCy, PostgreSQL, and Docker. My efforts in data analysis and visualization led to a substantial reduction in browser response times, from 30 seconds to just 2 seconds. By leveraging pandas and Matplotlib to identify and communicate performance bottlenecks to the engineering team, I played a crucial role in the successful release of a 2-year project following these improvements.
Tools: spaCy, Docker, Python, SQL, Natural Language Processing (NLP).
Senior Software Engineer, Karnov Group Denmark
April 2017 - March 2020 • Copenhagen, Denmark
I developed and delivered multiple applications, both frontend and backend, using JavaScript frameworks like ReactJS and NodeJS. I utilized Google Cloud Platform, specifically Kubernetes, to deploy Docker-based applications. Additionally, I contributed to the development of the Karnov Business Optimizer - a financial tool for accountants and managers.
Tools: Node.js, React.js, Kubernetes, Docker.
Founder and Software Engineer, Occam.ooo
November 2016 - March 2017 • Copenhagen, Denmark
I was the technical lead for a mobile development project for Mobilab in Germany, contributed to frontend development for Competencehouse in Denmark, and handled mobile development for WithDipp in the USA. I designed, engineered, and delivered a mobile app called Nu Events. Additionally, I managed company operations.
Software Engineer, colego.dk
July 2015 - October 2016 • Copenhagen, Denmark
I served as a frontend product developer and a Node.js developer for multiple backoffice applications. Starting in June 2016, I led both local and remote teams and re-implemented the internal Content Management System (CMS).
Tools: JavaScript, Node.js, AngularJS.
Software Engineer, Unwire
February 2014 - June 2015 • Copenhagen, Denmark
I was involved in several projects, including developing a payment administration web app for Nordea using Backbone, Jasmine for unit testing, and HTML5. I created a boilerplate project for all frontend development using AngularJS and developed a mailing list manager with Backbone, jQuery, HTML, and Bootstrap. Additionally, I worked on an Android/iOS hybrid app for Nordea using Backbone, JS/jQuery, HTML5, Ratchet, and Phonegap. Throughout these projects, I utilized Atlassian JIRA and followed Scrum methodology.
Data Analyst, Copenhagenize Design Co.
November 2012 - August 2013 • Copenhagen, Denmark
I analyzed data on cyclists’ behavior in Copenhagen, which led to collaborative publication featured in Business Insider. I also analyzed data for the Copenhagenize Index 2013, implemented the company’s website, and calculated the environmental impacts of 30km/h zones. Throughout these projects, I primarily worked with Excel and JavaScript.
Web Developer, Produtrónica, Lda.
2006 - 2012 • Sintra, Portugal
As a part-time freelance consultant, I developed a stock management tool and portfolio using PHP (WordPress) and JavaScript. I also utilized HTML and CSS for web design and Photoshop/Gimp for graphic editing.