Emploi Tunisie » Offres d'emploi à Tunis » Projet de fin d'étude: Application web basé sur java Wicket + REST

Projet de fin d'étude

  • Bertsch Innovation
  • 16 Rue Ibn Bassem Menzah 4
  • Il'y a 1 mois
Postes vacants:
3 postes ouverts
Type d'emploi désiré :
Temps plein
Niveau d'étude :
Etudiant
Langue :
Anglais
Genre :
Indifférent

Description de l'emploi

Bertsch Innovation is devoted to product-driven, multi-channel communication. At our European offices and at our engineering facility in Tunis, our 80+ employees develop and support our software, advise customers, and implement innovative solutions to optimize processes in the areas of product information management (PIM), digital asset management (DAM), electronic catalogues, mobile, and multi-channel publishing. We also provide customers and partners with expert training and top-quality professional services in all these areas. With more than 150 successfully implemented PIM/DAM projects and more than 600 e-CAT customers BERTSCH INNOVATION a leading medium-sized software companies in the European PIM/DAM/e-CAT/mobile sector.

Introduction to mediacockpit

mediacockpit is our JEE-based software product with standard modules for product information management (PIM) and digital asset management (DAM). mediacockpit is also an extremely flexible platform for developing custom solutions leveraging our flexible and extendable data model and system services in new and innovative ways. Nearly all aspects of the system are configurable to fit customer requirements and all data and services can be accessed via our standard HMI and documented, external APIs (REST, CMIS, Java, …).

Introduction to Chroma

Pragma’s flagship software Chroma provides applications for professional print shops and includes services to automatically collect, analyze and visualize company and machine data (CDA/MDA/PZA).

For more information, please visit:

https://www.pragma is.de/modules/custom/produkte/chroma.php?locale=en_GB .

 Proposal 1: Online Image Editor

 mediacockpit integrates with Adobe Creative Cloud via plugins for PhotoShop and other CC apps. For certain use cases, customers prefer to perform image edits directly in the web browser, then save the results as a download or as a variant or new version in mediacockpit. A common use case is to add fly outs or position numbers in an image. In the scope of this project, a new, browser-based, image editor will be integrated into our DAM module, mediasuite.

Objectives and goals

The objective is to leverage and integrate an existing online image editor component, for example https://ui.toast.com/tui-image-editor/ A brief selection and approval process (e.g. from license standpoint) for the component will be part of the project.

Once selected, it will be integrated into the existing DAM module in mediacockpit, the mediasuite. This could, for example, be done as a new entry in the “…” menu for a media asset:

The editor appears in a modal window allowing the user to edit the source file. Depending on the format of the original file (e.g. JPG, TIF, PNG, AI, PSD, SVG), it may be possible to edit the original asset directly or may be necessary to edit a rasterized rendition of this file.

After editing the asset resource, it should be possible to download the result, or also to update the existing source file, add as a derivate, create a new version or a new variant linked to the source asset, features provided by the existing REST API in mediacockpit.

As a stretch goal, the manipulations performed could be stored as a “processing profile” on the server and the processing replicated and result file updated if the source file changes (e.g. a new version is created). Depending on the technology selected for the editor, this may be possible using a javascript server (e.g. node.js) or Chromium embedding (to be evaluated in the project).

Backgrounds and prior work and guidance

Mediacockpit provides server-side image manipulation and application plugins for image editors already. The implementation of these tools can be referenced during implementation and integration steps. In addition, Bertsch Innovation will provide senior-level guidance in the areas of software engineering and graphic design to help successfully complete the project.

 Proposal 2: ML/AI and product recognition / product identification / product classi-fication (Cloud/AWS Focus using SageMill)

 Products in mediacockpit (and our desktop app e-proCAT) are typically structured in hierarchies: the levels in the hierarchy vary from customer to customer, but for example they can include: manufacturer, brand, product group, product, article, etc. The hierarchy should be designed to minimize the effort to organize and manage product information in the system.

In addition to this master hierarchy, there exist numerous standard classification systems for products like GS1 GPC, ETIM, UNSPSC, etc. to enable identification and exchange of product information. In the simplest scenario, a classification is simply a code that identifies what "type" of product it is, e.g. GPC 20000002: Butter (Shelf Stable). In other scenarios, a "class" can include and require specific information be maintained and sent with the product for the classification to be complete.

For certain industries and use cases, products are classified according to standardized classification systems. The task of classification activity is frequently manual and redundant: open product, look at specification text or image, search for similar (or already known) classification and assign classification to the product. Simple automation strategies for product identification and "pre-classification" exist but using ML and the computing power of the cloud, it should be possible to vastly improve on these rudimentary automations.

The goal is to leverage machine learning and artificial intelligence to reduce manual effort and increase the quality of automated product classification. This can be accomplished by leveraging existing classified product data from a variety of sources (data pools, that has already been classified e.g. GDSN, or customers with large data pools).

This reference data, already classified, will be analyzed and the textual data and other features are used to train the algorithm how to associate products with a given product class.

In addition, since mappings exist between certain classification systems (for example ecl@ss and ETIM), the algorithm can be further trained that if a product 123 is classified in System A as class Z, and class Z in System A is related to class YY in System B, then the reference data from class YY can leverage to train the algorithm to additionally identify products of class Z. This way, reference data from multiple systems (and possibly multiple languages) can be used to improve predictions.

The algorithm must learn with time, so it must provide possibilities to further (re)train the algorithm. For example, when manually classifying data in our software, the newly classified product and its features are sent to the service which updates the algorithm and its predictions improve over time.

The base algorithm will likely have to be based on a specific functional domain (and possibly language), because training all features for all domains would likely be prohibitive. Specific customer data (names, addresses, etc.) should be obfuscated/anonymized, so that the core model can be used for all clients. Features such as price might be sensible to obfuscate into specific buckets / ranges since the histogram of prices might be a powerful identifier for classification

Backgrounds and prior work and guidance

In 2019, we started initial work with AI and ML in mediacockpit for product classification. This work can be analyzed and certainly improved in the scope of this project. Since the initial work was complete, we have acquired additional reference data, and we hope this will improve the prediction results of the algorithm for more product classes, classification systems and languages. Bertsch Innovation will provide guidance in the areas of ML, working with SageMaker/AWS and our products and technologies in general to help successfully complete the project.

 Proposal 3: Mobile App(s) Strategy for Chroma

The Chroma application suite includes both desktop and web applications in a variety of technologies. A number of these applications include features would be well suited for mobile apps, for example:

Task management

Marketing assistant (desktop and web)

Meeting summaries (web)

In addition, Chroma has a hand scanner that is programmed for Windows CE and could be migrated to a modern application platform, e.g. iOS and Android.

Objectives and goals

We assume this project can be divided between two candidates. In the scope of the project, the candidates will:

- Together in remote sessions with Chroma developers, analyze the existing use cases and implementations

- Identify, interfaces between apps and Chroma

- Define the scope of the tasks per candidate (for example, one does Marketing assistant, another does Task / Project dashboard and/or the hand scanner app)

- Implement, test and document the mobile apps and – depending on scope definition – also the interfaces themselves.

Backgrounds and prior work and guidance

Chroma provides desktop (and to a lesser extent, web) applications that realize these features today. The implementation of these tools can be referenced during the project and implementation steps. In addition, we will will provide senior-level guidance in the areas of software engineering and graphic design to help successfully complete the project.

 Timetable and milestones

We propose starting on 01. February 2019 and are planning to complete the project by 30 June 2019 (a period of five months). A more detailed timeline can be set forth after acceptance.

Location

We are offering this project proposal through our office in Tunis, located in a representative villa near Stade El Menzah and the 10 Décembre Metro station. The candidate will work closely with the engineering team in Tunis the UI/UX and Chroma core developers in Germany.

Remuneration

The remuneration for the period shall be 300 TDN / month. An extension of this agreement, if necessary, to complete the project, is negotiable, but should be addressed no later than 15 May, 2019

Exigences de l'emploi

Compétences requises :

+Java, JavaScript et PL/SQL avec une bonne maîtrise des Technologies/Framework : J2EE, JSP ,HTML/CSS/SASS et JQuery.

+ Excellente compréhension des mécaniques de WebServices 

+Maitrise d’un Framework ou CMS est un plus 

+ Une bonne communication, maîtrise de l’ anglais

+ Maîtrise de l’allemand  est un grand atout. 

 Prière de Joindre votre  CV en Anglais ou en Allemand.

Date d'expiration

25/12/2019

A propos de Bertsch Innovation
Bertsch Innovation GmbH is one of the leading European providers of enterprise software for product information management, digital asset management and multichannel product communication. Our employees have all the key competences to help our customers achieve a...
Categories: Informatique

Offres d'emploi les plus recherchées

Casanova - Sousse, Tunisie
MANWORK - Tunis, Tunisie
Interfil - Tunis, Tunisie
Vocal-line - Tunis, Tunisie
Employeur?
S'inscrire pour un compte employeur et Postez vos offres!