Artificial Intelligence Engineer Nanodegree From Udacity
Written by Nikos Vaggalis
Thursday, 27 October 2016
In its pursuit to establish itself as a synonym to Training in AI, Udacity is looking to expand on the success of its Self Driving Car Engineer Nanodgree by launching a new offering with a more general outlook and widespread reach.
Due to the high bar set for entry to the Self Driving Car credential, a major portion of its hopeful candidates had to be turned away and this could provide them with a second chance. This doesn't mean that there's no barrier; there still is but lowered in such a way that the only qualification required is to be able to solve exercises of the following kind:
Given a set of N people on a social network how many possible "friend" relationships (i.e. pairs of distinct people) are there?
What will the following python code output?
s = 
while N > 0:
s.append(N % 3)
N = N / 3
buf = ""
while len(s) > 0:
buf += str(s.pop())
I have a fair coin and a two-headed coin. I choose one of them with equal probability and I flip it. Given that I flipped a heads, what is the probability that I chose the two-headed coin?
1+2+3+...+(N-1) = N(N-1)/2
1212, which is 50 written in base 3.
The underlying principle is that whoever manages to successfully solve these exercises, would also demonstrate a basic knowledge of linear algebra and calculus, have the ability to apply basic probability and statistics plus experience implementing computer science algorithms, and lastly, would know how to write a program in Python, all essential for tackling the course.
They're still somewhat formidable as entry level requirements, but still nothing compared to those the Self Driving Car nanodegree asks for by way of experience in Probability, Statistics and Machine Learning.
Compared to the Self Driving Car Engineer nanodegree, where those graduating would fill specific roles inside the tight boundaries of the self driving automotive industry, the likes of Vehicle Software Engineer and Sensor Fusion Engineer, Artificial Intelligence Engineer can approach the field of AI in a more holistic fashion with a much more widespread application. As such its graduates would be able to find work in the general AI domain, encompassing healthcare, finance, retail, media, advertising and education.
Nonetheless, specialization is not out of the question since students choose the path they would like to specialize in after acquiring the necessary general knowledge in the first term, therefore potentially developing themselves into System Software, Machine Learning, Predictive Analytics, NLP Software, Computer Vision and Automation Engineers. As such, the related classes cover topics such as Search and Optimization Logic, Natural Language Processing and Computer Vision, expecting students to dedicate up to 15 hours/week on the course's material.
Not surprisingly, interest in this opportunity went through the roof. With only 500 seats initially available, and applications already covering that number multiple times, the competition to reserve a spot in the workforce of the future, especially given the fact that graduates will be given preference by Nanodegree including Didi Chuxing and IBM Watson, is guaranteed to be fierce. For graduates of the 2011 AI class, Introduction to Artificial Intelligence, the fact that the instructor line up reunites Sebastian Thrun and Peter Norvig is another incentive to enrol.
For those not succeeding at first, do not despair since the 500 cap is not fixed in stone and looks to increase on each iteration.
Applications began on October 25th and are going to be open till November 14th. In contrast to the Self Driving Nanodegree 3-term duration, this one lasts 2 terms only, each term spanning three months and costing 800 dollars; that is $1600 in total.