How do I break into AI?
Don’t be afraid of AI or deeplearning! It’s exciting, even thrilling at times - and yes, frustrating at others. But it’s great fun. And it is for everybody. If you like, you can take me as an example...
I grew up in a world without computers. There wasn’t a single one at the school I went to. And even when I was 21, in my early days as a journalist, we wrote our articles on mechanical and later electrical typewriters. TippEx was a main tool of my profession. At the age of 26, I studied environmental sciences at UEA in Norwich, UK, and only Phd students had regular access to computers and advanced programs like “Excel”. I went back into journalism in 1994 and now there were indeed computers, but I struggled immensely to handle them.
I liked my job, worked for Bloomberg, Der Spiegel, lead the Companies & Markets department at Handelsblatt and finally became deputy editor of Focus, the second biggest news magazine in Germany. But I got increasingly bored with the same old discussions: Is Angela Merkel at the end of her political career? I’ve heard this so many times. And for quite a while I also felt that I was possibly missing something. There is very little space for stringent logical thinking in journalism. That is why, in 2007, I started to study Physics at the Technical University in Berlin, alongside my job as a correspondent covering the finance ministry for Der Spiegel. The financial crisis halted this phase. In 2012 I came, quite by chance, across a Java tutorial on the internet. I tried it, enjoyed it, and kept learning to program, mainly with Coursera. And for the first time in my life I had the feeling that programming was offering me the combination I had been looking for - creativity and stringent logical thinking. But I was still far off AI and machine learning. It seemed too complicated and I was probably a bit afraid of it.
In December 2016 I was visiting an auction house in Munich to research a story I was working on. During talks, the owner told me about a counterfeiter he had problems with. And I asked him: “Is there nothing digital to check the style of a painting?” He wasn’t aware of anything and I thought I’d try to develop something. I got 400 images from four painters, defined 52 hand-coded features and tackled the problem with Bayes theorem. It probably was sort of a naive Bayes classifier - and it reached 67 percent accuracy on predicting the correct artist. This encouraged me to improve it, but I quickly realized that I had to get into machine learning. I did Andrew Ngs first Coursera course (in Matlab) and when i applied it on my painter classifier, I realized the power and potential that neural networks have. I also did Andrews deeplearning specialisation and improved the result even further. With more than 10000 images and 15 painters I got an accuracy of 86 percent, which could (I am convinced) be improved even further by using, for example, more advanced networks.
About that time, roughly a year ago, at the age of 52, I thought about changing my profession completely and starting a new career in AI and deeplearning. In February I left Focus and worked as a freelance AI programmer. Since May this year I also work part time as a Senior AI Strategist for the appliedAI initiative at Unternehmertum, a German start-up incubator and innovation consultant.
I can maybe talk about how I started my journey and it might help you start yours -
One fine day, I came across an article from HBR which said 'Data Scientist: The Sexiest Job of the 21st Century'. This instilled in my mind and I loved how I started playing with data at my job as a business consultant. I switched jobs, worked with different clients, different data sources, and different teams but there was this one constant, DATA!
When you work with data-centric companies, you keep on hearing about advanced machine learning techniques. Terms that fascinate you, that scare you. To deal with this, I came back to taking online courses. I started with, yes you guessed it right, Machine Learning by Andrew Ng on Coursera. It got a little too much for me and I went astray for a few months. I had to pull myself back and I started with another course and a few more, before I stumbled upon The Analytics Edge on edX. This course broke a lot of myths that I had been carrying for years. This is precisely what got me started on my journey to discovering the power of data and analytics. I completed that course, enrolled in the Deep Learning Specialization by Andrew Ng, participated in Kaggle competitions, played with Twitter data, created wordclouds, made chatbots, gave talks in local conferences, joined meet-ups, contributed to open-source projects, wrote articles about my learnings and never stopped exploring.
Let me share the story of my personal path into AI, hoping that you can take clues from there. I studied business administration and global economics in college and worked afterwards as a business consultant for a global company. During my work there, I quickly noticed the power of data and fact-based thinking. It was also pretty obvious that the largest value add in our generation would come from digitalization. A quick look at the stock market shows you that the most valuable companies in the world are digital companies. Having the urge to truly understand digitalization and being capable of creating software products, I started to study towards my Master in Computer Science in part-time. After two years, I noticed that I needed to gain practical experience in a digital business as a programmer. I quit my consultant job and started to work as a working student. (Yes, the pay cut still hurts today). I first worked as a Java Engineer, before switching to the Data Science team. I wrote my Master Thesis on Data Science, while taking online courses from Andrew and other sources. After this year, I felt ready to enter the AI market and I’m happy that I found the perfect job for me to make an impact in the automotive industry.
I want you to take three things from this story. First, check your motivation! If you have a burning desire to use AI to better the world and create value for society, you’re on the right track. Second, continue to learn! Analyze your skills. How sufficient are you in Software Engineering, Statistics and Business? Once you know your skills, the next steps are where the fun starts. Don’t be afraid to take odd steps to advance your knowledge in the right direction. Bring your weaknesses to a minimum acceptable level. Take online courses on any of the three AI fields. Third, play your strengths! When applying for a job, I pounded on my business knowledge and how this would help me help the business. You also have plenty of strengths given your background, make them visible! Get your hygiene factors right through taking online courses on Software Engineering, Machine Learning, or business knowledge. Then build upon your strengths, and employers will be lucky to have you.
Lastly, let me offer some advice on the traits I’m looking for when hiring Data Scientists. When I’m interviewing Data Science candidates, I’m looking for three things: Background, strengths, and willingness to learn. I believe that a diverse team with diverse backgrounds is absolutely beneficial for any team. So don’t fret if you’re entering the Data Science field laterally, your background knowledge will make you valuable nevertheless. Then I’m looking which strengths you’re bringing and if those strengths will help the team. Third, I’m looking for your motivation to learn. Certificates of online courses are a huge plus, because they show extra-effort. Online courses tell me also that this candidate has valuable practical skills and is ready to help the team. Learning on the job will become ever more important, so if you show that you’re already learning new skills, you will remain valuable in any team.
Hi everyone. I started about a year and a half ago with no data science background, and only a little programming experience (though I’d studied physics & math in college). At that time I was working part-time, and was part-time at home with my kids. I wasn’t sure how possible it would be to return to work in a technical field after being away for around 8 years, but I decided that a good way would be to build up skills in AI.
I began with Andrew Ng’s Machine Learning course, and then worked through the whole Deep Learning sequence. I also learned Python (mainly using Jose Portilla’s courses on Udemy) and worked through the FastAI sequence and some Kaggle Competitions. Lastly I studied Sergey Levine’s Berkeley course on Reinforcement Learning, and Robert Sedgewick’s Princeton course on Algorithms (both on YouTube).
After all of this, I’m excited to now be a Fellow working on the Multiagent Team at OpenAI. I wish the very best to everyone here & want to encourage everyone -- even if you think right now that you don’t have the right background. It’s an exciting time in the field!
Hello everyone, I know friends that were not having any mathematical or coding skills, they have started reading articles about AI, and they took the "AI for everyone" course, and now they are telling stories about AI, I believe that all of us should adapt AI, today everything we need to know and to learn is available on the web, so for sure AI is so cool, the journey starts with one thing, it's passion!
Beyond passion, the most things helped me to know much about AI and adapt it, are taking Sir Andrew Ng's amazing courses on Coursera, reading various articles about AI, interacting on social media, and following all deeplearning.ai updates, your team is really awesome doing hard work to demystify everything we need to know, thank you!
Well, I don't think so you cannot break into AI. If you want to learn new application in Data Acience , AI, Data Analytics then you can join Data Analytics Course.