2. ALEX LILBURN
FORESEEING FAULTS WITHIN TELECOM NETWORKS.
Using machine learning to provide
services that fail less by creating
a model which can process time
series data to predict service faults
ahead of their occurrence.
I met with the application support
team at a major telecoms company to
hear about the methods employed in
the industry to pre-empt faults. Whilst
some patterns were being identified
the existing approach was extremely
labour intensive and often didn’t
provide enough information to
predict future incidents.
I saw analogies with some of the
latest machine learning techniques
I was studying at the time and so,
successfully, put myself forward to
apply some of these techniques to
the problem.
Task:
I was provided with around 9 months
of application incident logs to analyse.
There were two main sub-projects.
First, to identify and describe patterns,
dependencies and co-occurrences
so that the application support team
could better understand their data.
Second, to develop and test a number
of predictive models to be able to
pre-empt incidents, particularly
those which result in reduced service
provision to customers.
Review:
The project is still ongoing
but so far the data points
to improved fault detection
and prediction and we are
constructing a model for
implementation within
the business.
3. What makes this project unique?
We are taking a very different
approach to provide data insight
and create an accurate predictive
model that will be utilised within
the business to improve service
to real customers.
How involved in the
project were you overall?
Extremely. I have been responsible
for running all of the analysis
and developing the predictive
models from concept through to
implementation and evaluation.
The client has supported with
regards to providing domain
expertise, but it has otherwise
been a self led project.
What else are you working on?
I’m also working with Satalia, an
awesome optimisation business,
developing a tool which takes a
complex query of things you’d
like to see and do, and provides
you with a schedule which meets
your requirements as effectively
as possible.
Did you always know this was
the area you wanted to work in?
I have always had an interest in
technology and maths, but it took
a while for me to realise this was
my niche. I first studied Psychology
and then Law. I noticed the AI hype
and got really excited about the
impact I could make bringing these
techniques to novel areas. I turned
down a legal training contract offer,
applied for the Business Analytics
MSc course and the rest is history.
What has been the
highlight so far?
Data science wise, my career is still
fledgling, but completing v0.1 of
the scheduler for Satalia was really
exciting. It’s the first time I’ve put
together an end-to-end application,
and it works! I’m also very excited
about the impact my telecoms
project could have.
What advice would you
give your 18 year old self?
Think differently. Grab opportunities
with both hands and get stuck in!
What is changing in engineering
and why is it important?
It has never been easier to learn
to code and incredibly powerful
tools are becoming much easier
to use thanks to UX and clever
API design. It will bring greater
diversity to the industry allowing
people to spend more time actually
creating. In terms of innovation and
productivity, it’s a game changer.
////
Q&A
WE SAT DOWN WITH ALEX AND ASKED HIM A FEW
QUESTIONS ABOUT HIS PROJECT AND ASK WHAT HE
THINKS THE FUTURE HOLDS FOR HIMSELF.
“It’s never
been easier
to learn to
code...it’s a
game changer”