Development of the cloud based tracking, theft prediction and theft prevention services for bicycles within Kissmybike project. Risk Terrain Modeling approach was followed to build predictive models for cities based on bicycle thefts open data.Tech stack: Google Cloud Platform (cloud functions, compute engine, app engine, Firebase, cloud storage, BigQuery, Cloud Pub/Sub, IoT Core, Maps API), Cordova, jQuery Mobile, iOS, Android, Java, JavaScript.Development of micromobility optimization platform which studies existing city infrastructure and patterns of bicycle usage in the city to suggest adjustments to the infrastructure to meet the real needs of cyclists.Tech stack: Google Cloud Platform, Jupyter, papermill, pandas, geopandas, scikit-learn, statsmodel, OSRM, OSMnx, Shapely, mplleaflet, Matplotlib, Python.

Development of algorithms for accurate matching customer location information and Point of Sale data within RetailerIN project. RetailerIN is an advanced in-store analytics solution able to turn location information into actionable insights to improve processes and optimise store’s profitability.Tech stack: Jupyter, papermill, pandas, NumPy, SciPy, scikit-learn, statsmodel, sqlalchemy, xlrd, Matplotlib, Python.

Development of a recommender system for GreenPanda project that provides suggestions on experiences to be carried out based on the user profile information and experience logs.Tech stack: Docker, PredictionIO, Spark, HBase, Elasticsearch, Scala.Development of backend subsystem for collecting Wi-Fi sensor data, indoor positioning and data processing for in-store analytics computation.Tech stack: Docker, Kong, MQTT, Kafka, MySQL, Spark SQL, Livy, Yii PHP Framework.

Coordinating development of semantic search subsystem for Entitypedia entity repository.Tech stack: Lucene, Nutch, GATE, WordNet, OpenNLP, PostgreSQL, PostGIS, Hibernate, Spring, Swagger, JavaCC, JMeter, Java.