08/27/2024 updated
100 % available
Co-founder, Principal Data Scientist, Principal Data Consultant, Data Scientist
Paralimni, Cyprus
Worldwide
2013Geographic Information SystemsApplication Programming Interfaces (APIs)Amazon S3Data AnalysisOptical Character Recognition (OCR)ArcGIS (Software)ArcMapBig DataBing AdsSoftware QualityDatabasesData IntegrationData MiningData VisualizationData Warehousing
Redshift, Amazon S3, API, ArcGIS, ArcMap, Big Data, Bing Ads, Analytics, Data Analytics, Data Analysis, Data Integration, Data Mining techniques, Data Mining, Data Stream, Data visualization, Data Warehousing, databases, geospatial data, Google Ads, user interaction, user interface, Jenkins, Looker, Matlab, Machine Learning, matplotlib, object oriented, Optical Character Recognition, Oracle, ORACLE DATABASE, PL/SQL, pandas, PowerBI, PySpark, Python, PyTorch, R, R Programming, recommender systems, SQL, Snowflake, code quality, transfer learning, unsupervised learning, XGBoost
Experience
Professional experience10+ years
Freelancer since2017
Languages
GermanNative speakerEnglishGood
Project history
Successfully completed an urgent project ahead of schedule, meeting the agreed model
performance in terms of efficiency and accuracy, and surpassing expectations regarding code
quality:
* Implemented an scalable object oriented solution integrating Optical Character
Recognition and Image Processing techniques to automate the extraction and validation of
key information from documents
* Developed of an API designed for end-user interaction and integration
performance in terms of efficiency and accuracy, and surpassing expectations regarding code
quality:
* Implemented an scalable object oriented solution integrating Optical Character
Recognition and Image Processing techniques to automate the extraction and validation of
key information from documents
* Developed of an API designed for end-user interaction and integration
Main results:
* Online Marketing optimization: Implemented end-to-end an XGBoost stacking based model to
predict the potential value of new leads, enabling Target ROAS smart bidding in Google
Ads Campaigns, observing an increase in company ROAS of >15%
* Sales Optimization: Implemented end-to-end an algorithm which dynamically adapts spend
for the different products/destinations based on the sales agents calendar availabilities
reducing sales collapses >90% and presumably increasing the ROAS per lead and decreasing
the average lead time-to-call
* Designed and implemented the user interface for the stakeholders to control the
parameters of the previous algorithm
* Designed the BI dashboards for the stakeholders to monitor the previous algorithm
* Mentored junior team members in best practices for data science methodologies, betting
for a culture of continuous improvement and agile quality peer reviewing
* Online Marketing optimization: Implemented end-to-end an XGBoost stacking based model to
predict the potential value of new leads, enabling Target ROAS smart bidding in Google
Ads Campaigns, observing an increase in company ROAS of >15%
* Sales Optimization: Implemented end-to-end an algorithm which dynamically adapts spend
for the different products/destinations based on the sales agents calendar availabilities
reducing sales collapses >90% and presumably increasing the ROAS per lead and decreasing
the average lead time-to-call
* Designed and implemented the user interface for the stakeholders to control the
parameters of the previous algorithm
* Designed the BI dashboards for the stakeholders to monitor the previous algorithm
* Mentored junior team members in best practices for data science methodologies, betting
for a culture of continuous improvement and agile quality peer reviewing
Advised for the development of the Data Science area in the first stages of the company:
* Established the features summarising the users web activity (Necessary pillar for many
others Machine Learning projects)
* Implemented a Machine Learning model to predict probability of purchase of a lead based
on the historic web activity, automating the train set, target and validation set
generation following a proper time series methodology
* Performed a segmentation of the users and defined stability checks for the derived
segments
* Image processing: Performed transfer learning techniques for image classification, number
plate blurring and logo substitution
* Defined validation metrics to measure the recommender systems performance
* Carried out ad-hoc descriptive analytics to support product and design decisions on a
regular basis
* Proposed and supervised of our junior team members projects
* Established the features summarising the users web activity (Necessary pillar for many
others Machine Learning projects)
* Implemented a Machine Learning model to predict probability of purchase of a lead based
on the historic web activity, automating the train set, target and validation set
generation following a proper time series methodology
* Performed a segmentation of the users and defined stability checks for the derived
segments
* Image processing: Performed transfer learning techniques for image classification, number
plate blurring and logo substitution
* Defined validation metrics to measure the recommender systems performance
* Carried out ad-hoc descriptive analytics to support product and design decisions on a
regular basis
* Proposed and supervised of our junior team members projects