Of­fer 399 out of 421 from 11/06/19, 15:35


Smart Steel Tech­no­lo­gies GmbH - Soft­ware­ent­wick­lung

Smart Steel Tech­no­lo­gies tack­les the last effi­ci­ency fron­tier in steel manu­fac­tu­ring through Arti­fi­cial Intel­li­gence. We offer ready-to-use AI soft­ware pro­ducts that improve pro­duc­tivity and qua­lity in all steel manu­fac­tu­ring pro­ces­ses. Our world class Arti­fi­cial Intel­li­gence experts and pro­cess experts ensure that our pro­jects deli­ver per­ma­nent cost savings and reach break-even after 12 mon­ths.

Machine Lear­ning Working Stu­dent (m/f/d) - Indus­trial Data Sci­ence and Soft­ware


At Smart Steel Tech­no­lo­gies, we are buil­ding a one-of-a-kind cross-disci­pli­nary team, unit­ing the worlds of steel and machine lear­ning. Our team has come tog­e­ther from a diverse set of aca­de­mic and pro­fes­sio­nal back­grounds: soft­ware deve­lop­ment, com­pu­ter sci­ence, mathe­ma­tics, phy­sics, and metall­urgy. Our goal is to leverage this exper­tise through trust and coope­ra­tion. We want to fos­ter an envi­ron­ment where team mem­bers have fre­quent oppor­tu­nities to learn from each other in fields that they are not (yet) experts.

The machine lear­ning and data sci­ence we employ comes in several varie­ties: image pro­ces­sing, signal pro­ces­sing, clas­si­cal super­vi­sed mode­ling, unsu­per­vi­sed methods, and fun­da­men­tal sta­tis­tics and pro­ba­bi­lity. As a machine lear­ning working stu­dent, your spe­ci­fic exper­tise as well as your moti­va­tion would be the major fac­tors in deter­mi­ning your day-to-day work.

We are gra­te­ful to stand on the shoul­ders of the open source giants. We use NumPy-fla­vo­red Python3 as the basis for our deve­lop­ment, taking advan­tage of the avai­la­bi­lity of libra­ries desi­gned for both soft­ware engi­nee­ring and data sci­ence, ran­ging from Flask to SciPy, SQL­Al­chemy to Keras, Jupy­ter to Pan­das, and many more. When Python’s time or space com­ple­xity beco­mes a cons­traint, we turn to modern C++. When we require cli­ent-side brow­ser-based inter­ac­tivity, we turn to Java­Script.


  • Apply modern machine lear­ning methods to opti­mi­za­tion and qua­lity pro­blems in indus­trial pro­ces­ses
  • Tune and opti­mize neural net­works, both bes­poke and pre-archi­tec­ted; shal­low, deep and recur­rent
  • Deve­lop and pro­vi­sion soft­ware for online lear­ning and pre­dic­tion in indus­trial sys­tems

Er­war­te­te Qua­li­fi­ka­tio­nen:


  • Bache­lor’s or Mas­ter’s in Com­pu­ter Sci­ence, Sta­tis­tics, Mathe­ma­tics, Phy­sics or Engi­nee­ring
  • Expe­ri­ence in working with raw and unst­ruc­tu­red indus­trial data sets
  • Expe­ri­ence in soft­ware engi­nee­ring or ope­ra­ti­ons, espe­ci­ally in a full-stack deve­lop­ment envi­ron­ment
  • Expe­ri­ence with modern web tech­no­lo­gies
  • Flu­ency in Eng­lisch, Ger­man and Rus­sian a plus

Un­ser An­ge­bot:


  • Balance of auto­nomy and hier­ar­chy
  • Trans­pa­rent busi­ness stra­tegy
  • Mul­ti­di­sci­pli­nary team
  • Mea­ning­ful indus­trial pro­blems
  • Modern Machine Lear­ning
  • Sleek new office in Ber­lin
  • Edu­ca­tio­nal oppor­tu­nities

Hin­wei­se zur Be­wer­bung:

We look for­ward to recei­ving your app­li­ca­tion inclu­ding your salary expec­ta­ti­ons and avai­la­bi­lity at