Edgardo Di Nicola Carena
Head of Data Science at Abstract- Claim this Profile
Click to upgrade to our gold package
for the full feature experience.
Topline Score
Bio
Experience
-
Abstract
-
Italy
-
IT Services and IT Consulting
-
400 - 500 Employee
-
Head of Data Science
-
Jul 2019 - Present
I created a dedicated, peaceful and committed team for Data Science, managing all aspects such as careful selection of people who must be technically strong, true team players and full of soft skills. With this team we have successfully implemented complex data science projects in multinational organisations around the world.
-
-
Manager of New Technologies
-
2006 - Jul 2019
Abstract produces, sells and implements Lucy, the leading automated document flow system for the multi-format document and data capture market. Its customers include national and international companies in both the public and private sectors. Lucy can be integrated with the main ERP systems on the market, such as SAP. Abstract sells its products either directly or through partnerships with other service providers. Achievements: - Continuous 360-degree support to innovate and… Show more Abstract produces, sells and implements Lucy, the leading automated document flow system for the multi-format document and data capture market. Its customers include national and international companies in both the public and private sectors. Lucy can be integrated with the main ERP systems on the market, such as SAP. Abstract sells its products either directly or through partnerships with other service providers. Achievements: - Continuous 360-degree support to innovate and create new business opportunities. - Developed state-of-the-art techniques for automated extraction and validation of key information after applying OCR to unstructured paper documents. - Successfully innovated and outperformed the technological core of a leading software product (Lucy) for automated recognition and extraction of key information from paper invoices. The new engine, based on the current product specifications and with the addition of the most advanced auto-adaptive techniques (Fuzzy ARTMAP, neural networks, Bayesian inference, etc...). - Integrated the new engine into Abstract ‘s main software product, successfully replacing the previous engine and increasing the recognition rate from 45% to over 65%. - Implemented a new algorithm to identify the right supplier in more than 97% of cases, compared to less than 90% before (within the required timeframe of 30 days). - Designed and implemented a new engine for extracting product descriptions from noisy faxes that outperformed the existing engine, with correct transcription in over 50% of cases.
-
-
-
-
CTO
-
2000 - 2006
TNP solutions were based on my own patented concurrent application of neural networks and advanced statistical modeling. The technology combines conceptual analysis of content and patterns with user profiling and ranking. Meaning, user info, usage and geographical data are correlated in the conceptual hyperspace to build the user's profile and his degree of belonging to predefined or automatically discovered communities. Achievements: Managed R&D team for development of new… Show more TNP solutions were based on my own patented concurrent application of neural networks and advanced statistical modeling. The technology combines conceptual analysis of content and patterns with user profiling and ranking. Meaning, user info, usage and geographical data are correlated in the conceptual hyperspace to build the user's profile and his degree of belonging to predefined or automatically discovered communities. Achievements: Managed R&D team for development of new advanced software products. Developed and applied signal processing techniques for automatic data reconditioning and categorization, data mining, user collaboration, automatic user profiling and partial tagging on very large databases. Proved benefits of automatic data categorization and data mining modules, within collaborations with major companies. - Delivered own proprietary algorithms and technology for a new full internal web-based knowledge management system that is used by financial analysts to quickly find relevant documentation used to accomplish key financial studies and assessments. - Delivered own proprietary algorithms and technology to automatically real-time categorize and prioritize abuses like spam, phishing, hacking, pedophilia using various sources like warning messages from users and navigation logs. - Delivered own proprietary algorithms and technology to allow chain to better manage assistance through rationalization and retrieval of a huge amount of car mechanical failures and related successful fixes. - Delivered own proprietary algorithms and technology for categorization and semiautomatic publication of relevant internal documentation on both website and Intranet. Show less
-
-
-
University of Cambridge
-
United Kingdom
-
Research Services
-
700 & Above Employee
-
Research Associate
-
1998 - 2000
Cambridge, United Kingdom
-
-
Education
-
Politecnico di Milano
PhD, Electronic & Communication Engineering