By Devanshu Anand – 

No doubt Machine learning is changing the face of various industries but to make this

change ML engineers play a very important role. According to John Glover who is the
Head of Research at Aylien, Dublin does not have an average day but rather John and
his team mates have an average week depending on the project they are currently
working on. Their tasks include doing literature review, new product releases, building
and training the models, evaluation and at last deploying the models to build a
prototype. According to John, Back Propagation in Deep Learning is one of his favourite
algorithms as it helps him to do many of the tasks in his team, which also coincide the
problem solving skills as well. The hardest challenge in a person’s life while pursuing
PhD is to step in for something new or commencing for the first time. This can be a life
changing event or a backward step. In my opinion, being an ML engineer takes all of
your effort, skills, time period, and originality to create something denovo or continue
salvage. Researchers need to learn basic technical aspects, analyze machine learning
algorithms, identifying differences in data distribution, verifying data quality through
filtering, and defining validation strategies.

Average day of an ML engineer does take hours to compile the data sets, analyse them,
make changes according to the problem faced, even produce some new solutions to it in
between. ML engineers need to understand the entire ecosystem for which they are
designing. One of the most interesting things is there are virtually no fields in which
machine learning cannot be applied. To add on, my experience is kind of taking twists
and turns making me realise the worth and importance of ML and its network. During
my initial PhD days, I got an opportunity to take an interview and meet one of the great
industrialists, and knowing the facts about their work, how they work on such types of
algorithms and many more, which made me come back to the fact of being an ML

If I talk about the journey through all of these ups and downs, while doing 2 or 3 tasks at
a time, makes us feel the true potential of a human brain and the output it gives at the
end. I have observed, reading is one of the ultimate meta skills in Machine Learning, if
there was a better way of doing what I was doing, I could save time and effort by
learning it and implementing it. And there was a deadline approaching, then I could cut
short the material and carry on with my project. And the most difficult thing you can
come across, is communication of algorithms rather than technologically proven. To
conclude, as said by John, we do PhDs in this particular field, to research and search

Full interview with John Glover