Blätter-Navigation

Of­fer 294 out of 605 from 09/07/21, 15:51

logo

Fraun­ho­fer Hein­rich-Hertz-Insti­tut - For­schung

Fraun­ho­fer is Europe’s lar­gest app­li­ca­tion-ori­en­ted rese­arch orga­ni­za­tion. Our rese­arch efforts are gea­red ent­i­rely to people’s needs: health, secu­rity, com­mu­ni­ca­tion, energy and the envi­ron­ment. As a result, the work under­ta­ken by our rese­ar­chers and deve­lo­pers has a signi­fi­cant impact on people’s lives. We are crea­tive. We shape tech­no­logy. We design pro­ducts. We improve methods and tech­ni­ques. We open up new vis­tas.

Stu­dent Assi­stant* - Front­end deve­lo­per for health.aiau­dit.org

Work­ing field:

Towards health AI tech­no­logy that can be trusted for use in real-world app­li­ca­ti­ons

Modern AI sys­tems based on deep lear­ning, rein­force­ment lear­ning or hybrids the­reof con­sti­tute a fle­xi­ble, com­plex and often opa­que tech­no­logy. Limits to our under­stan­ding of an AI sys­tem’s beha­vior con­sti­tute risks for sys­tem fail­ure. Reli­ably mana­ging the risks asso­cia­ted with modern ML tech­no­logy remain an open sci­en­ti­fic and prac­ti­cal hurdle.

In order to remove these hurd­les, we envi­sion a frame­work for algo­rithm audi­t­ing and qua­lity con­trol along the ent­ire ML4H life cycle that will pro­vide a path towards the effec­tive and reli­able app­li­ca­tion of ML tools in health­care. For that pur­pose, we col­la­bo­rate in the ITU/WHO Focus Group on Arti­fi­cial Intel­li­gence for Health (FG-AI4H) to design methods, pro­ces­ses and stan­dar­di­za­tion con­tri­bu­ting towards AI tech­no­logy that can be trusted for use in real-world app­li­ca­ti­ons

We are working with open source frame­works such as Eva­lAI (https://health.aiaudit.org/) and MLflow to deve­lop solu­ti­ons for auto­ma­ted audi­t­ing, fede­r­a­ted audi­t­ing in remote teams and auto­ma­ted report crea­tion.

You can find out more about our work in the fol­lo­wing video and paper as well as our code repos and col­la­bo­ra­tion site.

Your pro­s­pec­tive respon­si­bi­li­ties:
  • Adapt the exis­ting Eva­lAI-Angu­larJS based front­end to the new user inter­face requi­re­ments
  • Front­end-user inter­face design using the Angu­larJS frame­work (for both plat­form admin and other user roles)
  • Front­end-manage­ment dash­board GUI design - for chal­len­ges, data­sets, sub­mit­ted solu­ti­ons, and bench­mar­king results and repor­ting lea­der­boards
  • Django REST API design - to estab­lish con­nec­tivity of front­end modu­les with other soft­ware modu­les/ packa­ges (e.g. API design for Angu­larJS - Django backend con­nec­tivity)
  • Com­mu­nity out­re­ach with the FG-AI4H topic groups

Require­ments:

  • Stu­dent* at a European uni­ver­sity (e.g. com­pu­ter sci­ence, engi­nee­ring). NOTE: We may also be able to offer con­tracts for non-EU stu­dents on a pro­ject basis. Please let us know in case this app­lies to you.
  • Because we aim at a lon­ger-term col­la­bo­ra­tion, you should be enrol­led as a stu­dent* for at least one year.

Skills that will help you do well:
  • Know­ledge in ML: model deve­lop­ment, deploy­ment and tes­ting
  • Know­ledge about Eva­lAI frame­work: soft­ware archi­tec­ture
  • Requi­red: Expe­ri­ence with Python, Node.js, Angu­larJS, HTML, CSS, Scrip­ting, Django, Docker and Git
  • Favor­abl: Expe­ri­ence with AWS and Ter­ra­form,
  • Good com­mand of writ­ten and spo­ken Eng­lish
  • As we are cur­r­ently working remo­tely, we espe­ci­ally value an inde­pen­dent and struc­tu­red work style

What we of­fer:

We work remo­tely as our pro­ject teams come from dif­fe­rent parts of the world. In a typi­cal week, we have one or two group mee­tings to sync up on work. We like to give all team mem­bers time and space to get work done in-bet­ween mee­tings and try to avoid micro-manage­ment. Nevertheless, we are there to sup­port each other and use slack, dis­cord as well as email for asyn­chro­nous com­mu­ni­ca­tion. We use git for code ver­sio­ning, Google docs and mark­down for docu­men­ta­tion and git­hub pro­jects for task manage­ment. For most of the heavy infra­st­ruc­ture (Docker regis­try, data­ba­ses, vir­tual machi­nes/ser­vers, task queu­ing, worker manage­ment) we use AWS pro­ducts. Every two or three mon­ths there is a large ple­nary mee­ting for the FG-AI4H. Usually the weeks prior can be a bit more stress­ful as we pre­pare to finish up on mile­stones for pre­sen­ta­tion. Next to the deve­lop­ment of methods and soft­ware, we also regu­larly sub­mit our work to sci­en­ti­fic jour­nals or con­fe­ren­ces for pre­sen­ta­tion and dis­cus­sion.

  • Prac­ti­cal expe­ri­ence in machine lear­ning and cloud com­pu­ting
  • A highly moti­va­ted, inter­na­tio­nal team
  • A mea­ning­ful task to build your AI/ML and soft­ware port­fo­lio
  • Fle­xi­ble working hours and excel­lent equip­ment
  • Super­vi­sion by expe­ri­en­ced prac­ti­tio­ners and sci­en­tists from renoma­ted com­pa­nies and rese­arch insti­tu­tes
  • Bet­ter pay­ment in com­pa­ri­son to uni­ver­sity

How to ap­ply:

Please sub­mit your app­li­ca­tion docu­ments (cur­rent cer­ti­fi­cate of enrolment, aca­de­mic degree cer­ti­fi­ca­tes, employ­ment refe­ren­ces, CV, etc.) via our recruit­ment por­tal. Your app­li­ca­tion can­not be con­si­de­red wit­hout the rele­vant sup­por­ting evi­dence.