You’re going to need a quicker computer. If you want to become part of the AI revolution, the Tortoise Intelligence team has produced a small starter pack.
What kit do I need?
Computer programming is centred around processing and translating information. Having a laptop that can process lots of information quickly is a good starting point.
Storage: data requires space, and you’re going to need a lot of it if you want to do data science and AI. So invest in a top-notch SSD (solid-state drive). Minimum memory for the SSD should be 256GB, as speed is critical.
Processors: 8GB RAM is recommended, but the higher the better. If you plan to handle more hardcore data, you can always move to cloud-based computing instead. Services like Amazon Web Service or Python Anywhere, a favourite with our editors, let you leave work running in the background without worrying about closing your laptop.
Screen size: should be 13” or higher, preferably 14-15”. Anything much bigger gets a bit too heavy to lug around.
Resolution: data involves a lot of visualising and plotting so a higher resolution screen is easier on the eyes.
Keyboard: programming involves lots of typing so your keyboard must be robust. Having a US-English keyboard will help with making shortcuts you’ll learn later down the line.
One member of the Tortoise Intelligence team has a US laptop but it’s set up as a British keyboard meaning their keys don’t quite match what they type… because mastering AI wasn’t hard enough.
A few laptop recommendations:
Which programming language to choose?
Python is the most popular computer programming language in data science and AI, and is the current language of choice taught in the majority of academic institutions. The language helps users build good coding techniques quickly, is easier to read, and has a wide range of features for working data.
The language’s popularity is in part down to the availability of a large number of specialised libraries, particularly in Artificial Intelligence, including PyTorch and TensorFlow.
Which degree should I study and where?
It’s now possible to get a BSc in Artificial Intelligence, but Maths, Computer Science and Physics remain popular options as data science jobs often advertise for these degrees. As for where to study, the Tortoise AI Index uses data from The Times’ World University Rankings.
For Computer Science:
University of Oxford, UK
– Entry requirements: A*AA with A* in Maths, Further Maths or Computer Science at A-level;
– Course length: three years (BA), four years (MCompSci);
- Course fees: £27,750 (three years) or £37,000 (three years).
California Institute of Technology, US
– Entry requirements: STEM extra-curricular, 4.0 in GPA, and a close to perfect SAT score;
- Course length: four years (Undergraduate Major in Computer Science);
- Course fees: tuition is around $46,000 per year.
University of Cambridge, UK
– Entry requirements: A*A*A;
– Course length: three years (BA) four years (MEng);
– Course fees: £27,750 (three years) or £37,000 (four years).
The course ranking was the same for Mathematics and Statistics.
DeepMind Computer Science Scholarships are awarded to students wishing to pursue a Master’s degree at Oxford University. Applicants must be a resident in the UK and belong to one or more of the following groups: female, BAME, or from households with a traditionally low progression to higher education.
Can I teach myself?
Top universities aren’t the only option, though. There are hundreds of free, open-access courses online, also known as ‘MOOC(s)’. Popular providers include edX, Coursera, DataCamp and Codeacademy, which have a plethora of options to choose from for getting started in the AI world. These courses are a more affordable and flexible way of learning. Courses include Artificial Intelligence, Machine Learning and Python, to name a few.
One course recommended by the Intelligence team is the Coursera course on Machine Learning, led by Stanford University. Taking around 56 hours to complete, you’ll be in good company with the 2.7 million people already enrolled. The course itself is free, but if you’d like a certificate as proof that you’ve mastered machine learning, you’ll need to pay a £61 fee.
For a more technical and specific course that needs some pre-requisite background knowledge, Stanford’s CS224N Natural Language Processing with Deep Learning is available on YouTube for free. This is the course one of our data scientists is currently doing.
Learning new coding tricks doesn’t have to be structured, though. Google is your best friend – and searching a problem you’re facing with your code typically takes you to a site called Stack Overflow. It’s the Yahoo Answers of the coding world, with professionals and amateurs alike helping each other to squash bugs and exchange tips. If you’ve got an issue, simply post a question, along with your code, and 99 per cent of the time someone who’s encountered the same challenge will point you in the right direction.
CV credits and competitions
To stand out from the crowd, it’s always useful to have a few extra credits at the bottom of your CV. Adding links to your GitHub profile, a code hosting platform, will help; as well as to your Kaggle and Twitter accounts. These act like a portfolio where you can show recruiters the work you’ve done.
A few AI-based challenges:
- Try out Kaggle’s Machine Learning from Disaster challenge. Here, you can build models that can predict who survived the Titanic;
- Check out freeCodeCamp’s tutorial on predicting sentiment of tweets;
- Or get your web-scrapers out and classify your favourite movies from IMDB with DataQuest.
What are the rewards?
Most likely a job, and a good one at that.
There’s a huge choice of sectors to work in. Data scientists often go on to work in defence, healthcare, robotics and pharmaceuticals.
A 2018 LinkedIn study found that, based on data from its platform, the US currently has a shortage of 151,717 people with data science skills. Jobs platform Indeed found a similar skills gap, reporting a 30 per cent increase in demand for data scientists year-on-year, and a 348 per cent increase since 2013.
So, the few people with the right skillset are in high demand.
Most entry level positions in AI have job titles including Data Scientist, Research Scientist, and Data Engineer. According to the jobs site Glassdoor, a Data Scientist for AI can expect to earn £55,708 on average in London.
In the United States, the annual salary survey by Dice, a tech recruiter, found that the role of a Data Scientist carries an average salary of $106,000 a year, increasing to over $120,000 in large tech-based cities such as San Francisco.