No, I don’t just ask ChatGPT to tell me
Data Science moves fast.
Like, really fast.
Five years ago, if you said you were interested in AI that roughly translated to something between having a PhD in Quantum Mathematics and having an “I heart π” tattoo plastered across your back. Now my mum sends me ChatGPT memes and you can literally buy picture books that will teach neural networks to your babies.
The days of blagging your way through AI conversations are well and truly over. Nowadays, you actually have to know stuff. So how do you stay up to date with what’s going on in the world of AI and Data Science?
While there’s already lots of great advice out there, I’ve found that many of the strategies recommended by others don’t work for me. As a Data Scientist working in industry, I have very limited time to spend on research, and it wouldn’t make sense to spend all that time pouring over academic papers and technical release docs. I need sources that are short, snappy and applied — sources that help me understand how to make use of new technologies in my day to day work.
In this guide, I’ll summarise the sources that have worked well for me. My aim is to highlight those that are helpful from an applied perspective (not just a business or research perspective). If you’re a Data Scientist or aspiring Data Scientist, my hope is that these sources will help you stay relevant and spot opportunities for innovation in your day to day work.
And with that, let’s get into it.
Lots of Data Science teams have blogs — my personal favourites being the blogs of Netflix, Tripadvisor, Duolingo, Meta and Spotify. The thing I like about these is that they’re super practical. Whereas a lot of academic research papers focus on the theoretical aspects of Data Science and Machine Learning, companies’ blogs tend to highlight how Data Science is being used in practice to tackle real-world problems.
As a Data Scientist working in industry, I find these really helpful. I don’t have to sift through reams of theoretical jargon or decipher hundreds of greyscale flowcharts. I can just get right to the heart of the issue and understand how Data Science is used to tackle the kinds of problems I’m facing in my day-to-day job as a Data Scientist.
The other day, for instance, I was looking for some information on A/B testing: a common framework in Data Science that I personally have little experience with. I found lots of OK-ish articles online, and then stumbled upon this fantastic series on Netflix’s tech blog. Not only did it answer my questions about A/B testing; it was also packed full of practical examples of how A/B testing is used at Netflix. This was immensely helpful for understanding not just the theory but also the practice.
One of the main reasons I started writing on Medium was to help me learn.
Not to teach or to market; to learn.
It’s often said that you remember 10% of what you read, 20% of what you see and a whopping 95% of what you teach. The accuracy of these specific numbers is by-the-by: the point is that if you solely rely on reading others’ blogs, you’re unlikely to extract the full value or retain all the information you’re reading. Starting a Data Science blog is an excellent way to consolidate this knowledge and reinforce the things that you’re learning. It forces you to think through things step-by-step and helps you gain a deep understanding of the subject matter.
If the idea of writing online sounds a bit out of your comfort zone, don’t worry. I felt exactly the same! But ask yourself: what’s the worst that could happen? In my case, I worked out that the worst possible outcome was a bit of light-hearted teasing from my colleagues. On balance, I reckoned that I could handle that, and that it would be more than worth it given what I’d gain from the experience of writing.
If you’re not sure what to write about, you might enjoy this article I wrote last week about how to come up with Data Science project ideas:
Or, take a look at some of the suggestions on Towards Data Science’s FAQs page: that was the place I looked when I was searching for initial ideas.
Reading and writing blogs are great ways to stay up to date with the latest technical developments in Data Science and AI, but if you want to keep up with the business side of things, I’d recommend subscribing to AI-focused newsletters which keep track of startups, acquisitions and research trends.
My personal favourite is TLDR AI Newsletter: they send one daily email summarising the key news stories in the industry, and it’s much more focused than their more established tech newsletter. Alternatives include The Download by MIT Technology Review and The Batch by DeepLearning.AI. I love newsletters like these because they take me literally 2 minutes to read and help me quickly understand what’s going on outside of my small corner of the AI world.
DAIR.AI maintains a fantastic GitHub repo which is updated weekly with a list of 10 new machine learning papers. For each paper, DAIR provides a brief summary and link to a tweet explaining more about the findings of the paper. If you don’t want to ‘Watch’ the GitHub repo or subscribe to email updates, you can also follow DAIR.AI on Twitter.
Personally, I find that these short summaries are much more “actionable” than the alerts you get from services like Google Scholar or arXiv. When setting alerts on Google Scholar and arXiv, you can quickly become inundated with notifications and, if you’re anything like me, end up not engaging at all due to information overload. The great thing about DAIR.AI is that they perform the legwork for you, filtering to the most interesting and cutting-edge papers and giving you the option to either read more or quickly discard and move on.
Two Minute Papers is a YouTube channel which does exactly what it says on the tin. Each week, two new videos are uploaded to the channel, each of which aims to distill the findings from a recent research paper, many of which focus on AI. As I write this, there are 495 videos in their AI and Deep Learning playlist.
Subscribing to this channel is a great way to keep abreast of the latest developments in AI research. I particularly love their classic “OpenAI Plays Hide and Seek” video, but honestly there are so many fantastic ones it’s hard to choose.
Two other channels that I particularly like are StatQuest with Josh Starmer and 3Blue1Brown. I love these channels because they explain statistics and machine learning concepts in a really visual and understandable way, without assuming loads of prior knowledge. While these channels are more famous for their introductory courses, they do also publish explainers of cutting edge topics in machine learning, for example this recent video on convolutions by 3Blue1Brown:
Lots of organisations host free webinars where they’ll talk through the latest innovations in Data Science and AI. Personally, I’m a big fan of these because booking onto a webinar forces me to make time for learning and development, which is really helpful for ensuring that I have protected time to spend on keeping up to date.
For example, if you use a cloud database system like Google BigQuery or AWS RDS in your day-to-day work, you might benefit from going along to a webinar hosted by Google or AWS which talks about how to get the most out of these tools. I recently attended one of these (a fantastic BigQuery webinar) on the theme of optimising your SQL code to reduce costs and query runtimes.
If you’re anything like me, it can be easy to get lost in the technical details of many AI press releases and technical documentation pages. Twitter is a great place to read more candid opinions and tips from experts. I’d particularly recommend following people like Yann LeCun, Timnit Gebru, Geoffrey Hinton, Andrew Ng, and Christopher Manning. AI researchers like these use Twitter to share quick-fire updates on their work and share useful tidbits that don’t always make it into scientific papers.
The main reason I like Twitter, however, is because it’s a great way to follow Data Scientists and AI practitioners working in industry (not just research). People like Chris Albon, Jay Alammar and Cassie Kozyrkof work on the kinds of problems that are very relevant to the day-to-day life of a Data Scientist working in industry, and following people like them is a great way to get an idea of what other Data Science orgs are working on.
One more thing: Twitter has a better sense of humour than arXiv.
Staying up to date with the latest trends in AI and Data Science is absolutely vital if you’re trying to build a career in this space.
In my experience, however, I’ve found it easy to get sucked into the minutia of my specific tasks and lose track of what’s going on in the industry as a whole. In this article, my aim has been to highlight the strategies which have helped me stay up to date with new trends and spot opportunities for innovation in my day to day work. If you think I’ve missed any gems, please let me know in the comments — I’d love to hear about what’s worked for you.
No, I don’t just ask ChatGPT to tell me
Data Science moves fast.
Like, really fast.
Five years ago, if you said you were interested in AI that roughly translated to something between having a PhD in Quantum Mathematics and having an “I heart π” tattoo plastered across your back. Now my mum sends me ChatGPT memes and you can literally buy picture books that will teach neural networks to your babies.
The days of blagging your way through AI conversations are well and truly over. Nowadays, you actually have to know stuff. So how do you stay up to date with what’s going on in the world of AI and Data Science?
While there’s already lots of great advice out there, I’ve found that many of the strategies recommended by others don’t work for me. As a Data Scientist working in industry, I have very limited time to spend on research, and it wouldn’t make sense to spend all that time pouring over academic papers and technical release docs. I need sources that are short, snappy and applied — sources that help me understand how to make use of new technologies in my day to day work.
In this guide, I’ll summarise the sources that have worked well for me. My aim is to highlight those that are helpful from an applied perspective (not just a business or research perspective). If you’re a Data Scientist or aspiring Data Scientist, my hope is that these sources will help you stay relevant and spot opportunities for innovation in your day to day work.
And with that, let’s get into it.
Lots of Data Science teams have blogs — my personal favourites being the blogs of Netflix, Tripadvisor, Duolingo, Meta and Spotify. The thing I like about these is that they’re super practical. Whereas a lot of academic research papers focus on the theoretical aspects of Data Science and Machine Learning, companies’ blogs tend to highlight how Data Science is being used in practice to tackle real-world problems.
As a Data Scientist working in industry, I find these really helpful. I don’t have to sift through reams of theoretical jargon or decipher hundreds of greyscale flowcharts. I can just get right to the heart of the issue and understand how Data Science is used to tackle the kinds of problems I’m facing in my day-to-day job as a Data Scientist.
The other day, for instance, I was looking for some information on A/B testing: a common framework in Data Science that I personally have little experience with. I found lots of OK-ish articles online, and then stumbled upon this fantastic series on Netflix’s tech blog. Not only did it answer my questions about A/B testing; it was also packed full of practical examples of how A/B testing is used at Netflix. This was immensely helpful for understanding not just the theory but also the practice.
One of the main reasons I started writing on Medium was to help me learn.
Not to teach or to market; to learn.
It’s often said that you remember 10% of what you read, 20% of what you see and a whopping 95% of what you teach. The accuracy of these specific numbers is by-the-by: the point is that if you solely rely on reading others’ blogs, you’re unlikely to extract the full value or retain all the information you’re reading. Starting a Data Science blog is an excellent way to consolidate this knowledge and reinforce the things that you’re learning. It forces you to think through things step-by-step and helps you gain a deep understanding of the subject matter.
If the idea of writing online sounds a bit out of your comfort zone, don’t worry. I felt exactly the same! But ask yourself: what’s the worst that could happen? In my case, I worked out that the worst possible outcome was a bit of light-hearted teasing from my colleagues. On balance, I reckoned that I could handle that, and that it would be more than worth it given what I’d gain from the experience of writing.
If you’re not sure what to write about, you might enjoy this article I wrote last week about how to come up with Data Science project ideas:
Or, take a look at some of the suggestions on Towards Data Science’s FAQs page: that was the place I looked when I was searching for initial ideas.
Reading and writing blogs are great ways to stay up to date with the latest technical developments in Data Science and AI, but if you want to keep up with the business side of things, I’d recommend subscribing to AI-focused newsletters which keep track of startups, acquisitions and research trends.
My personal favourite is TLDR AI Newsletter: they send one daily email summarising the key news stories in the industry, and it’s much more focused than their more established tech newsletter. Alternatives include The Download by MIT Technology Review and The Batch by DeepLearning.AI. I love newsletters like these because they take me literally 2 minutes to read and help me quickly understand what’s going on outside of my small corner of the AI world.
DAIR.AI maintains a fantastic GitHub repo which is updated weekly with a list of 10 new machine learning papers. For each paper, DAIR provides a brief summary and link to a tweet explaining more about the findings of the paper. If you don’t want to ‘Watch’ the GitHub repo or subscribe to email updates, you can also follow DAIR.AI on Twitter.
Personally, I find that these short summaries are much more “actionable” than the alerts you get from services like Google Scholar or arXiv. When setting alerts on Google Scholar and arXiv, you can quickly become inundated with notifications and, if you’re anything like me, end up not engaging at all due to information overload. The great thing about DAIR.AI is that they perform the legwork for you, filtering to the most interesting and cutting-edge papers and giving you the option to either read more or quickly discard and move on.
Two Minute Papers is a YouTube channel which does exactly what it says on the tin. Each week, two new videos are uploaded to the channel, each of which aims to distill the findings from a recent research paper, many of which focus on AI. As I write this, there are 495 videos in their AI and Deep Learning playlist.
Subscribing to this channel is a great way to keep abreast of the latest developments in AI research. I particularly love their classic “OpenAI Plays Hide and Seek” video, but honestly there are so many fantastic ones it’s hard to choose.
Two other channels that I particularly like are StatQuest with Josh Starmer and 3Blue1Brown. I love these channels because they explain statistics and machine learning concepts in a really visual and understandable way, without assuming loads of prior knowledge. While these channels are more famous for their introductory courses, they do also publish explainers of cutting edge topics in machine learning, for example this recent video on convolutions by 3Blue1Brown:
Lots of organisations host free webinars where they’ll talk through the latest innovations in Data Science and AI. Personally, I’m a big fan of these because booking onto a webinar forces me to make time for learning and development, which is really helpful for ensuring that I have protected time to spend on keeping up to date.
For example, if you use a cloud database system like Google BigQuery or AWS RDS in your day-to-day work, you might benefit from going along to a webinar hosted by Google or AWS which talks about how to get the most out of these tools. I recently attended one of these (a fantastic BigQuery webinar) on the theme of optimising your SQL code to reduce costs and query runtimes.
If you’re anything like me, it can be easy to get lost in the technical details of many AI press releases and technical documentation pages. Twitter is a great place to read more candid opinions and tips from experts. I’d particularly recommend following people like Yann LeCun, Timnit Gebru, Geoffrey Hinton, Andrew Ng, and Christopher Manning. AI researchers like these use Twitter to share quick-fire updates on their work and share useful tidbits that don’t always make it into scientific papers.
The main reason I like Twitter, however, is because it’s a great way to follow Data Scientists and AI practitioners working in industry (not just research). People like Chris Albon, Jay Alammar and Cassie Kozyrkof work on the kinds of problems that are very relevant to the day-to-day life of a Data Scientist working in industry, and following people like them is a great way to get an idea of what other Data Science orgs are working on.
One more thing: Twitter has a better sense of humour than arXiv.
Staying up to date with the latest trends in AI and Data Science is absolutely vital if you’re trying to build a career in this space.
In my experience, however, I’ve found it easy to get sucked into the minutia of my specific tasks and lose track of what’s going on in the industry as a whole. In this article, my aim has been to highlight the strategies which have helped me stay up to date with new trends and spot opportunities for innovation in my day to day work. If you think I’ve missed any gems, please let me know in the comments — I’d love to hear about what’s worked for you.