Introduce yourself here

miwojc · January 30, 2020, 2:19pm

for your twitter we would need to know your twitter handle thanks!
so if you are logged in on twitter click on your profile and then you can copy and paste url

ChristophNeuner · January 30, 2020, 6:23pm

Oh yeah I did not proofread my text.
Thanks for notifying me
I changed it.

go_go_gadget · January 30, 2020, 6:27pm

Hi, I’m Laura and I’m currently a mathematics lecturer, hoping to transition into a data science role in industry this year. I worked through Part 1 this past fall, but felt like most of it went over my head, so I’m eager to revise with the new version. I’m particularly excited to read that perhaps the content from the ML course will make an appearance!

For my contribution to the community, I think I’m going to set a goal of asking all the stupid questions. I always tell my students to ask any questions they think of, because if one person is confused, there’s a good chance someone else is, too. So I’m going to take one for the team, and just ask. Fortunately, the fast.ai community is friendly and helpful!

I blogged about my experiences with Lessons 1-6 of Part 1 of the 2018 version, and those can be found here, for anyone looking for a preview of a newbie’s experience with the course.

travis · January 30, 2020, 7:07pm

Hi everyone! I started with fast.ai in 2017. I’m still very much a novice, but I love the fast.ai teaching philosophy, and I’m trying to spread the knowledge among the high school students I work with each day. I am an assistant principal at a STEM school near Savannah, GA. I have competed in a few Kaggle competitions, with two top 10%-finishes. I am currently recruiting students to join me in the upcoming NCAA tournament challenge.

jeremy · January 30, 2020, 7:23pm

That would be the most amazingly helpful thing I can think of.

Also, if you see answers to questions that are more complex than they need be, clarifications or clarifying questions are very helpful.

Pak · January 30, 2020, 8:29pm

Hi, my name is Pavel, I’m from Moscow (Russia). I am a gamedesigner.
I’ve been following fastai course since it’s first version. One of russian tech influencers tweeted about fastai’s awesome learning technique and, damn, he was right
Since the first inception I can’t recommend enough your course to everyone who asks me about ML/NN (and sometimes I do it even if they don’t ask ) as your way of teaching (from the practical results to core theory) is muuuch more efficient in terms of motivation, at least for me personally.

As I said I am a gamedesingner and I’m looking forward to find a way to use fastai in my profession (after all this summer of AI wouldn’t be so shiny without game industry and it’s cheap graphic cards ). Not much so far and I’ve ended up with tabular data analysis with NN (as we have no deficit of tables in games).
As Jeremy suggested in Introduction to Machine Learning course I’ve made an implementation of Feature Importance and Partial Dependence with fastai library (version 0.7 I think). And then I’ve used it to have fun in trying to analyse (european) Football transfers. And as it appear fastai is not only capable of predicting players’ transfer value as good as people (including professionals from transfermarkt.de), but it also can help to catch some interesting insights about this market.

I am looking forward to start a new course and learn how the library and fastai techniques evolved through these 1.5 years.

And Jeremy (or maybe someone from the previous private groups can tell) I have a question. In the initial post you’ve mention that it will be a live stream. Would it be possible to watch it some hours (days) later as time difference will hardly let be watch it online?
Github

muellerzr · January 30, 2020, 8:33pm

I’ll add that your tabular resources I constantly go back and check anything I do (feature importance, etc) , and I’m still learning from those notebooks!!!

You should be able to if it’s a YouTube live stream, I’ve noticed mine go live the second after the stream is done

ramon · January 30, 2020, 9:13pm

Hi Everyone!

My name is Ramon Hollands from the Netherlands. Freelance software developer and really excited about fast.ai and being part of this community! I got on the fast.ai train in 2018 and am loving the ride ever since. After watching the lessons over and over and struggling to understand, I finally started implementing some models in production. To understand things more deeply I’ve started to implement the arcface paper from scratch.

Looking forward to more deeplearning fun!

ab_ai · January 30, 2020, 9:26pm

Hello all,

My name is Aaron. I’m responsible for Digital Transformation at my company and I am an evangelist for deep learning adoption in the company. I’ve taken Part 1 twice and I’m looking forward to some new and old content in the next session.

Twitter: @ab_was_taken

liberus · January 30, 2020, 10:17pm

Hello everyone, I’m Andrii!

I’m from Ukraine, working as a tech lead in one of the local outsourcing companies.
My main focus is Computer Vision but I’d like to further learn about other fields too.
I did a couple of interesting projects, one of which was action recognition using 3D CNN so feel free to contact me regarding this or similar topics.

Anyways, right now I’m getting into academia and hope that fastai will help me with my own research and will push my inspiration further. Also, depending on the time available, I’ll try to translate fastai to Ukrainian, hopefully, this will help newcomers to get in DL.

waydegg · January 30, 2020, 11:45pm

Hello everyone,

I’m Wayde, a Full Stack ML Engineer from San Diego, California. I’m excited to be a part of the new Part 1 course as it was great going through it last year with everyone, and it looks like the new and improved Fastai library is going to kick ass . I’m hoping to use the things we learn over the coming months towards deploying ML applications for personal projects and (at some point soon, hopefully) an employer .

If anyone is hiring I’m looking for work! Check me out at my personal website that has links to all my socials and some write ups.

Anyways, I’m excited to work with everyone again this year and to really learn how to leverage the power that is practical deep learning with Fastai!

jeremy · January 30, 2020, 11:48pm

Yes you can. You will miss out on the opportunity to ask questions during the class, of course. I’d strongly suggest trying to watch it no later than the next day, because the conversation online moves fast!

eunosM3 · January 31, 2020, 2:26am

Hi, my name is Matthew. I live in the US and I aspire to be useful in the DL field…someday. In the meantime, I’ll keep my day job as an institutional investor.

Johnpal · January 31, 2020, 2:42am

Hi everyone.

My name is John Garcia, from Colombia.

I am an information / cybersecurity consultant that started to learn ML/DL four months ago.

I recently took the 2019 part I course, and learned a lot in a short time, mainly because fast.ai library is so great that people like me, who coded 15 years ago at the University, may implement things since the first week!

My main topic of interest is trustworthy AI, so I am learning about fairness, accountability, explainable AI, and of course using my domain expertise of privacy and cyber security to deal with these kind of risks. I would love helping companies and people to build trustworthy AI systems.

I am thinking how to use fast.ai v2 to integrate 3rd party tools dealing with the topics I talked above, and also as a tool to evaluate other metrics to detect bias and create explainable systems.

amqdn · January 31, 2020, 5:21am

Hi, everyone –

My name is Andrew Nguyen, currently working as an ML Engineer, managing infrastructure and orchestration for production models on healthcare data.

I started the fast.ai courses almost two years ago, and in that time, I went from being a novice programmer and researcher to landing a spot in Fellowship.AI, and now I work with Launchpad.AI and their clients on AI products, thanks in large part to the course.

Always excited about the unconventional, creative thinking that’s characteristic of fast.ai that leads to big wins in the AI space, and the advances in the field now are as exciting as ever!

mhanan · January 31, 2020, 6:33am

My name is Matanya (@HananMatanya), I work as an NLP developer in a startup called DigitalOwl (and completing first degree in BSc). In the company, we do medical documents processing in the field of insurance.
I owe to the fast.ai community a lot!

The project with which I was accepted for work was Hebrew ULMFIT.
I created a learning group at my place of residence on fast.ai v3 part 1. The group completed the course successfully!
I think fast.ai v3 part 2 is one of the best practical courses on the web today in Deep Learning. A lot of my work environment is built on this course. It allows you to think and perform almost any experiment I want.
Anyone who deals with advances NLP with fast.ai is welcome to contact
To decentralize our experiments we use FastEc2.

I was very happy to receive the email and very excited about the course.

bhoomit · January 31, 2020, 7:58am

Hello everyone,

Bhoomit (@bhoomitt)
Work @ GoIbibo/MakeMyTrip, India
Building an internal Chatbot Platform
Using ULMFiT & fastai on production for intent classification for about a year

Learned most of what I know about NLP with V1. Looking forward to V4.

butchland · January 31, 2020, 8:43am

Hi,
I’m Butch Landingin, a software developer from the Philippines.

I first started my ML journey years ago with Andrew Ng’s ML course on Coursera, which in contrast to the fastai approach, uses a bottom up approach – starting with matrix algebra then building it up to understand more advanced concepts like forward propagation, back propagation and gradient descent .

While Andrew is great at explaining complicated mathematical concepts and making them easy to understand, it’s a hard course to master especially when you want to start applying deep learning to real world problems.

In contrast, when I first took the fastai course a year ago, I was very excited to be able to apply what I learned right away (and get good results as well!) just after the first few lessons.

I am excited to join this next iteration of the course and its improved fastai v2 library and would like to thank Jeremy, Rachel and Sylvain for building a fantastic course, a great library and an inclusive community!

– Butch
twitter : @butchland

AndreaPi · January 31, 2020, 10:31am

Hi! My name is Andrea Panizza, I’m a Senior Staff Data Scientist at Baker Hughes and on Twitter I’m @unsorsodicorda. I got interested in Statistics years ago while working as an Aero Design Engineer for turbomachinery, then I started learning about Bayesian Inference, Statistical Learning…and finally in 2016 I got on the Deep Learning train As I’m mostly self-taught in these topics, I followed quite a few MOOCs in the past.
I wanted to start again by following the fast.ai MOOC and this time I tried to start a study group in my hometown, without success. So this time I’m going to do it the classic way, i.e., with you folks as my wonderful remote fellow students!

BTW, let me thank @jeremy and all the fast.ai people not only for this great library, but also for:

an excellent getting-started guide to GCP
finally convincing me to open my blog It’s still in beta, so to speak, but finally after a long procrastination, Jeremy’s excellent template for GitHub pages convinced me to start!

vova · January 31, 2020, 11:40am

My name is Vova, I’m from St Petersburg, Russia.

I work as software engineer since long time (like since ~2004) and was never involved in ML/DL in my day job role. Though, I’m enthusiastic about learning new stuff in general, and in ML in particular.

Last year I’ve enjoyed Swift for TensorFlow part of fast.ai course and even took some part in its development by creating SwiftCV library, contributing to S4TF itself, and doing a fun project with S4TF.

This year I suddenly for myself became PyTorch/OpenMined grant recipient for working on Federated Learning open-source project with PySyft team! PySyft is privacy-preserving deep learning framework based on PyTorch.

Privacy is a relatively new (yet important) thing in DL, and I’d love to hear more on it in the new 2020 part, @jeremy

My main takeaway from fast.ai is that you don’t have to have PhD to do things in ML. Thank you for that!
Feel free to connect with me in Github or Linkedin.