Acorn Newsbot
Junior Member
- Joined
- Jan 28, 2006
- Posts
- 22,477
- Reaction score
- 115
“You know what, I didn’t even know how to really use a computer until I was about 15,” says Nominet’s associate data scientist, Ilyas Ali. “I wasn’t interested in the subject as a teenager – apart from when I wanted to beat my friends at fantasy football”.
“I was really into figuring out which players were doing well each week, and I knew that a friend of mine was using machine learning (ML) to predict the scores. He was winning every year, and I learned coding so that I could beat him,” he adds. “A year later, I went to university and thought I should study data science – and the rest is history.”
Ilyas’ route into data science might not be conventional, but it is a great example of how diverse real-world applications can make the field more appealing and accessible. Coding doesn’t need to be some kind of dark art, and it’s not something that you need to dedicate your early years to, to study it at a higher level.
We sat down with Ilyas to mark National Coding Week. We got his thoughts and advice on how anyone can make a start in data science, just like he did.
The right resources
“The biggest tool I had, surprisingly, was YouTube. That and speaking to people that are doing the same work that you are, that can help explain it to you,” advises Ilyas. “You can get a lot of support from users on GitHub that post solutions to different problems you may have too.”
But don’t be tempted to just copy and paste the solutions you find: “This won’t be as helpful. When you’re learning, you need to understand what each little thing does in the code. In that sense, there are actual ML websites that produce detailed guides that help walk you through the code from beginning to end – exactly what’s happening when you use it, and different methods to employ.”
Go the extra mile
If you’re studying data science, you’re often focused on the minimum work that you need to do to pass your courses – naturally. But one thing that Ilyas would recommend to progress quicker is doing challenges and quizzes in your spare time.
“When you’re looking for jobs in data science, you’re often asked whether you have written ML models before that are published. And the answer is usually no, because you were studying the whole time,” reflects Ilyas.
“A lot of people don’t realise you can do challenges on Kaggle or GitHub, write it on code in your spare time and upload it to GitHub. Then when someone asks you, “Have you ever done work on a machine learning model before?” you can say, “Yes, I have! And here are all my solutions that I’ve posted online”. I found that out much later in life and I would recommend it to anyone.”
Stay curious
One thing that Ilyas loves about his job is the process of developing theories, testing them, and then adjusting his approach based on the results. He’s found that having an iterative, curious mindset is crucial for data science.
“Right now, I am working on a ML model that predicts whether every domain in Nominet’s registry is going to renew or not. I have to build a model that can predict, with a certain amount of accuracy, if something renews or not, and I have free reign of exactly how I want to do that – there’s no actual right way or wrong way of doing it. You need creativity to come up with, and try to answer, the questions that you pose yourself.
“If data science was a strictly math problem, there would be option A, B, and C, and I would need to tick each one off the list. It wouldn’t be as effective as the method we use at Nominet, which is coming up with theories, answering questions, testing adding and removing things, and using that iterative process in collaboration with other people.”
On this, he also notes that sometimes a simple suggestion from someone else would spark a new idea or approach that he hadn’t considered before: “Being open to input from others and exploring different avenues can help you uncover unexpected findings.”
Don’t seek perfection
He finishes on what he feels is some of the most important advice he was given in his career. “It’s completely fine to make mistakes, that’s why we have peer reviews – it’s completely normal,” assures Ilyas.
“And the second is more data science specific – when I first started in my career, I was chasing 100% accuracy on my models. But accuracy is not what you are striving for. It’s about trying to make a model that is easy to understand can answer a business-wide issue.
“Think about any ML model as a child that’s learning a language for the first time. If you teach it one word over and over again, it will be perfect at only that word. Whereas if you teach it a variety of things, it will make mistakes. If you keep trying to get it more accurate, you’re probably hurting the model more than you are helping it.
“If the model is perfect, how will it evolve? How will it learn? What can I understand from it? I think this way of thinking can be applied to your own work too.”
Ilyas’s own journey to becoming a data scientist shows us that there is no single “right” way into the profession. What matters most is a passion for problem-solving, an eagerness to explore new techniques, and the humility to continuously expand your own knowledge.
To mark this year’s National Coding Week, we we’ve shown that coding and data science isn’t all about being a technological whizz, and that everyone can have a go. As the field continues to advance, it will be these human qualities – the drive to learn, the willingness to experiment, and the spirit of mentorship – that will truly unlock its potential.
Are you interested in developing your digital skills? Nominet has funded Click Start – a nationwide training programme, developed by the Institute of Coding, to offer FREE skills and training to more than 26,500 learners across the country who may not have previously had opportunities.
Courses cover a spectrum of skills, from technical skills such as basic coding or specific tools, to personal and professional development. Find out more here: https://instituteofcoding.org/click-start/
The post From fantasy sports to forecasting domains: How to carve your own path in the tech industry appeared first on Nominet.
Continue reading...
“I was really into figuring out which players were doing well each week, and I knew that a friend of mine was using machine learning (ML) to predict the scores. He was winning every year, and I learned coding so that I could beat him,” he adds. “A year later, I went to university and thought I should study data science – and the rest is history.”
Ilyas’ route into data science might not be conventional, but it is a great example of how diverse real-world applications can make the field more appealing and accessible. Coding doesn’t need to be some kind of dark art, and it’s not something that you need to dedicate your early years to, to study it at a higher level.
We sat down with Ilyas to mark National Coding Week. We got his thoughts and advice on how anyone can make a start in data science, just like he did.
The right resources
“The biggest tool I had, surprisingly, was YouTube. That and speaking to people that are doing the same work that you are, that can help explain it to you,” advises Ilyas. “You can get a lot of support from users on GitHub that post solutions to different problems you may have too.”
But don’t be tempted to just copy and paste the solutions you find: “This won’t be as helpful. When you’re learning, you need to understand what each little thing does in the code. In that sense, there are actual ML websites that produce detailed guides that help walk you through the code from beginning to end – exactly what’s happening when you use it, and different methods to employ.”
Go the extra mile
If you’re studying data science, you’re often focused on the minimum work that you need to do to pass your courses – naturally. But one thing that Ilyas would recommend to progress quicker is doing challenges and quizzes in your spare time.
“When you’re looking for jobs in data science, you’re often asked whether you have written ML models before that are published. And the answer is usually no, because you were studying the whole time,” reflects Ilyas.
“A lot of people don’t realise you can do challenges on Kaggle or GitHub, write it on code in your spare time and upload it to GitHub. Then when someone asks you, “Have you ever done work on a machine learning model before?” you can say, “Yes, I have! And here are all my solutions that I’ve posted online”. I found that out much later in life and I would recommend it to anyone.”
Stay curious
One thing that Ilyas loves about his job is the process of developing theories, testing them, and then adjusting his approach based on the results. He’s found that having an iterative, curious mindset is crucial for data science.
“Right now, I am working on a ML model that predicts whether every domain in Nominet’s registry is going to renew or not. I have to build a model that can predict, with a certain amount of accuracy, if something renews or not, and I have free reign of exactly how I want to do that – there’s no actual right way or wrong way of doing it. You need creativity to come up with, and try to answer, the questions that you pose yourself.
“If data science was a strictly math problem, there would be option A, B, and C, and I would need to tick each one off the list. It wouldn’t be as effective as the method we use at Nominet, which is coming up with theories, answering questions, testing adding and removing things, and using that iterative process in collaboration with other people.”
On this, he also notes that sometimes a simple suggestion from someone else would spark a new idea or approach that he hadn’t considered before: “Being open to input from others and exploring different avenues can help you uncover unexpected findings.”
Don’t seek perfection
He finishes on what he feels is some of the most important advice he was given in his career. “It’s completely fine to make mistakes, that’s why we have peer reviews – it’s completely normal,” assures Ilyas.
“And the second is more data science specific – when I first started in my career, I was chasing 100% accuracy on my models. But accuracy is not what you are striving for. It’s about trying to make a model that is easy to understand can answer a business-wide issue.
“Think about any ML model as a child that’s learning a language for the first time. If you teach it one word over and over again, it will be perfect at only that word. Whereas if you teach it a variety of things, it will make mistakes. If you keep trying to get it more accurate, you’re probably hurting the model more than you are helping it.
“If the model is perfect, how will it evolve? How will it learn? What can I understand from it? I think this way of thinking can be applied to your own work too.”
Ilyas’s own journey to becoming a data scientist shows us that there is no single “right” way into the profession. What matters most is a passion for problem-solving, an eagerness to explore new techniques, and the humility to continuously expand your own knowledge.
To mark this year’s National Coding Week, we we’ve shown that coding and data science isn’t all about being a technological whizz, and that everyone can have a go. As the field continues to advance, it will be these human qualities – the drive to learn, the willingness to experiment, and the spirit of mentorship – that will truly unlock its potential.
Are you interested in developing your digital skills? Nominet has funded Click Start – a nationwide training programme, developed by the Institute of Coding, to offer FREE skills and training to more than 26,500 learners across the country who may not have previously had opportunities.
Courses cover a spectrum of skills, from technical skills such as basic coding or specific tools, to personal and professional development. Find out more here: https://instituteofcoding.org/click-start/
The post From fantasy sports to forecasting domains: How to carve your own path in the tech industry appeared first on Nominet.
Continue reading...
