Søren Johansen Email & Phone Number

Copenhagen, Capital Region, Denmark

Copenhagen, Capital Region, Denmark

Data capability team

Capital Region, Denmark

Copenhagen Area, Denmark

Copenhagen Area, Capital Region, Denmark

Using python nlp packages spaCy, textacy and fasttext for text mining on electronic patient recordsThe last year I have spend a lot of time with Deep Learning particular Keras. Experience with jetson tx2 and opencv (including contribution packages). Jetson tx2 implementing YOLO (You only look once) model. Object tracking opencv on Raspberry PI with Intel Movidius (Deep learning acceleration). Defined a Deep learning model for power plant performance optimization (Keras). Next step is to use reinforcement learning to see if we can improve performance (mse). POC Identification of text on engineering drawings (CAD/CAM). I build a python script that used opencv(preprocessing) and Azure Cognitive Services OCR. Databricks. Running hackathons for Data Science teams to use Databricks for Data Science colloboration and as last -resort to do machine learning on big data. Azure Machine Learning Services(AMLS) is python package that can orchestrate Azure Storage and Compute context (Spark, GPU). Hackathons cover the entire Model DevOps wheel from data to model deployment (Docker & Kubernetes). Model Monitoring. Deep learning model for identification of faces and number plates used BatchAI to do inference on many a large set of images. Azure Machine Learning Services was used to define a GPU compute context in Azure. Microsoft Machine Learning and Data Science Conference 2016 (Internal). Accepted Paper "Spatial Statistics".Deep Learning Institute. Certified trainer 2017.

Global Black Belt Technology Solutions Professional Advanced Analytics:Focus areas: Microsoft R Server(Scalable R functions) & HDInsight Spark. Many Models scenarios. Microsoft Deep Learning framework CNTK. Solving technical questions(R Server) and liase with Product Management and peers. Labs & Tutorials on Microsoft R Server and Azure Advanced Analytics (Machine Learning, HDInsight and Power BI) for partners and customers.Workshops on Deep Learning (Microsoft CNTK & Tensorflow, Caffe). Microsoft Data Science Tools(IDE): Visual Studio with R Tools & SQL Server.Hackatons & POCs: Insurance Fraud, Energy Forecasting & Many Models. Use of Azure Advanced Analytics (ML Azure + HDInsight + PowerBI).Parallelization alternatives in R: foreach, rxExec on Spark. Presentations at conferences on Advanced Analytics and Deep Learning. Supporting Product Managers visits in the Nordics: Datalake, R Server & Cortana Intelligence Suite.

Nordic

My main focus is to use advanced statistical and data mining techniques to solve business challenges such as: • Airline analytics• Churn• Cross sell/up sell• Loyality card data • Fraud detection• Sensor Data and Classical Insurance Pricing• Forecasting of magazine inventory levelDiscussions with customers on Big Data and how external data can be used to enhanced predictive models.Analytics with Hadoop:- Hortonworks sandbox- Cloudera LiveExplaining how the added-value of predictive analytics is when the models get integrated in customer facing systems in relationship to the holistic concept Decision Factory.

Nordics

Currently my main areas of focus is: • Solvency 2 assessments (gap analysis)• Insurance pricing(with GLM)• P&C reserving• Model lifecycle managementSpecial interests:• Statistical and data mining software• Model factory development• Visualization of data • Realtime scoring • Decision Management• Big data (and how to use external data to enhance prediction power)

My main focus was to use advanced statistical and data mining techniques to solve business challenges such as: • Churn• Cross sell/up sell• Loyality card data • Fraud detection• Pricing decisions• Forecasting of magazine inventory levelDiscussions with customers on Big Data and how external data can be used to enhanced predictive models was also part of the workshops. Explaining how the added-value of predictive analytics is when the models get integrated in customer facing systems.

Project managementIAS 39 Impariment loss chargeIAS 39 Hedge effectivenessFair value of options Business Intelligence for actuarial departmentsEnergy Risk