Saúl Vargas Email & Phone Number

Tallinn, Ee

Tallinn, Harjumaa, Estonia

As a member of the verification automation team, our mission is to continuously improve our verification engine so that it makes highly accurate decisions with minimal human intervention. As such, I have created or improved several key components of this engine using computer vision, data analysis, algorithms, etc. These are some of the projects I have worked on:- Input quality assessment: I have created services that are able to assess the quality of pictures taken by users (document images, selfies, etc.) to ensure optimal performance of our engine. If you ever use Veriff and we ask you to take a picture again, now you know who to blame.- Barcode scanning: identify documents all over the world still rely on good-old barcodes (linear, QR codes, PDF417) to store user data. I have improved our ability to detect and scan barcodes in identity documents, and added support for hundreds of barcodes types all over the world.- System design: developing machine learning models is always exciting, but being able to use them in an efficient and sustainable manner is what really makes me tick. I have spent considerable time refactoring key parts of our biometrics-related logic to ensure consistency, reusability and maintainability.Moreover, I have also been busy with:- Engineering standards: I have led initiatives across the whole engineering division to improve our Python development practices, namely the adoption of static type checking and the usage of lockfiles (poetry, PDM).- A/B testing: I have developed tooling to simplify the creation, execution and analysis of A/B tests.

Tallinn, Harjumaa, Estonia

• Healthy buildings (sensor-based insights for improving comfort in office spaces): iterated through the product from pilot validation to first public release.• MLOps: created a command line tool for easier use of AWS SageMaker with custom Docker images and metrics and artifacts tracking in MLflow.• Sensor-based legionella compliance: improved the accuracy of the machine learning model behind it.• Others: setting up data collection pipelines (weather data), setting up infrastructure (Terraform), deployment of machine learning models (MLFlow, TensorFlow, AWS SageMaker).

Tallin, Condado De Harju, Estonia

• Order conversion and customer churn: lead a project to estimate the impact of driver quality on those, the findings were used to prioritise initiatives at the company level.• OCR (OpenCV, Tesseract): developed algorithms to extract data from scanned documents, which saved thousands of hours of manual labour and allowed the company to comply with the regulations in one of their biggest markets.

London, United Kingdom

• Recommendation algorithms (personalised product ranking and you may also like product recommendations): improved significantly the quality of our recommendations by implementing state of the art algorithms (TensorFlow) which increased sales and engagement (A/B testing).• MLOps: lead several initiatives to adopt modern Machine Learning tools (Python, TensorFlow, Azure Machine Learning).• Others: team roadmap definition, mentoring and management, promotion of our work in external events.

London, United Kingdom

• Mendeley Suggest (recommendation of research papers): improved the relevance of recommendations by taking into account user’s recent interests and previous interactions with recommendations.• ScienceDirect’s “related articles”: developed a Spark-based implementation of an item-to-item collaborative filtering algorithm.

Glasgow, United Kingdom

• Participation in the SUPER FP7 European project: leveraging social media for crises and emergencies.• Academic research in social media analysis, sentiment detection, community tracking and recommender systems for venue suggestion and efficiency.

Madrid Area, Spain

• Participation in several national research projects: RIM3, MAVIR, Predict, Voxpopuli.• Academic research towards PhD.• Undergraduate lab teaching: programming, data structures, algorithms.

Barcelona Area, Spain

Usage of Yahoo's query log data and distributed computing infrastructure to propose new ways of assisting users formulate queries in mobile devices. Working with Roi Blanco and Peter Mika.

Barcelona Area, Spain

Definition of novel approaches for providing more diverse content in videoand music online platforms. Working with Linas Baltrunas and Alexandros Karatzoglou.

Glasgow, United Kingdom

Research on new solutions to alleviate the ambiguity and under-specification of short queries in web search engines.

Phd, Computer Science And Telecommunication

Novelty and Diversity Evaluation and Enhancement in Recommender Systems

Master'S Degree, Computer Engineering & Telecommunication

Exchange Student, Mathematics And Computer Science

Licentiate Degree, Computer Science

Licentiate Degree, Mathematics