Introduce yourself here

[Brainkite (Antonin Sumner)]

Thanks! This is amazing!

[suvash (Suvash Thapaliya)]

@czechcheck

I am mostly interested in applying ULMFit approach to genomic sequences

I’m a bit rusty right away with DL n fastai libs, but let’s start collaborating on this one at some point during the course.

[Redknight (Federico Lois)]

Hi everyone,

My name is Federico Lois and I am from Argentina (South America). You may have heard of us, probably not because of the beautiful places but because after Venezuela and Zimbabwe we are the country with the highest monetary inflation of the world (https://en.wikipedia.org/wiki/List_of_countries_by_inflation_rate) other possibility is because of Maradona, Messi and Manu Ginobili. Because of all the turmoil I am currently closing down 2 Argentinian companies during the following months and probably start working as a free agent in Machine Learning and/or Performance.

TLDR: I like to build stuff.

Since the very beginning, I have been using GPUs to do computing, it suffices to say, that my first GPU program involved register combiners (that gives up my age too :D). For the last 10 years, I have focused on applied research on Machine Learning and Performance. I was the guy that did all the ground research for the design of the storage engine for RavenDB 4.0 Transactional Storage Engine, also implemented and optimize it until release. The outcome was very well received with performance improvements of over 60x on average. My first involvement in Machine Learning happened back in 2013 (way before deep learning was the state of the art in the field) building a bulk photo face recognition engine for Intel. I got involved in deep learning shortly after when we were discussing the ‘next steps’ before Intel restructuring happened and the whole division I worked for as a contractor disappear into thin air.

On late 2017 I started to work on a deep learning-powered high volatility asset (cofff, cryptos) trading just for fun. Fast forward to today I have deep reinforcement learning-based engine that is being tested by trading desks and built some novel (unpublished) methods to deal with the typical noise and general poor behavior of financial time series.

I am also a recurrent speaker at the Russian DotNext conference where I focus mostly on low-level performance optimization and had talked also in a few other venues like NDC and QCON.

I hope I didn’t bore you with the details.

EDIT: Feel free to contact me also on www.twitter.com/federicolois or via https://www.linkedin.com/in/federicolois/

[sparalic (rustic)]

Hi All,
I’m excited to be part of this amazing community!!
My name is Sparkle and talk about transitions. I am originally a Chemical Engineer who made a switch to get PhD in Biomedical Engineering and now Data Science.
I am currently working as a Senior Data Scientist in a Biotech company here in the Bay Area. My work spans digital health to now medical imaging in the Ophthalmology space, so I look forward to apply what I’ve learned to my research.

I have learned a ton from this community and wanted to give back to the community by helping folks who are new to ML feel welcomed at some of the top ML conferences. As such, I helped co-organized the New in ML workshop at last year’s NeurIPS and we are currently working on creating the same footprint at this year’s ICML conference in Austria this July.

I’m looking to learning more with you all!!!

[dhruvt93 (Dhruv Thakur)]

Hi everybody. I’m Dhruv. I work as a backend/ML engineer at the Austrian startup TourRadar. fastai has been the best resource for me to learn about practical deep learning over the last one and a half year. I absolutely love what @jeremy, @rachel, and @sgugger have created at fastai. I just want to sincerely say a word of thanks, y’ll are awesome
find me here: twitter, blog

[sahilk1610 (Sahil KARKHANIS)]

Hello,

I am Sahil from Germany, I work for a consulting firm as a Machine Learning Engineer and have been a part of this wonderful fastai journey for 2 years now. Really excited to get my hands on the new fastai library.

A special note, maybe for the more experienced guys in the field or also to the admin it would be great if we spend like 30 40 mins in the course to actually know how deep learning models go into production specially the ones which are using the fastai library. A good know how of this is really beneficial in the long run.

Regards,
Sahil

[ulat]

Hi folks!
I am very happy to be part of this exclusive community! Thanks to Jeremy for developing this great DL-library!
Like many of you I started my AI/ML/DL Journey years ago with the first coursera class from Andrew NG. I startet DL with TF and Keras and switched to pyTorch with the arise of fastai.
I am working as a technical consultant for DL and college lecturer in Austria.

[miko]

Hi everyone. I’m Michelangelo (Miko for everyone but my mum). I am an Italian mathematician (abstract discrete math) turned product manager turned data scientist currently living in London UK and working at a utilities company, where I am helping to build the data function in the midst of a huge tech transformation process.

Things I have done in my current role (in the last year or so):

1. Use ULMFIT to build a call transcription classifier. I was thrown into this project when a lot of time had been invested into it with no clear objective. So, after months of attempts I managed to stall it as still had no clear way of bringing value. In the meanwhile, my little model trained on a single 1080ti in a sixth of the time and up to slightly better accuracies compared to Google’s AutoML

2. Updated the fastai ML course (some of the libraries used were out of date) and used it as the basis of a mini course to upskill software engineers who wanted to do so

3. Used the code developed above to trim down a 1000+ attribute table down to a mere 10 to perform customer loyalty. It took me less than a couple of days and the results on the validation set are frankly astonishing (and there’s no data leak, I promise)

4. Used a lot of the material published by Rachel to start establishing a good, ethical data science culture among stakeholders before we actually start hiring other data scientist and put models into production

On the side, I am also advising an AI startup and thinking of migrating their codebase to fastai2, as their research would benefit a lot from it.

Really looking forward to this!

[jeremy (Jeremy Howard)]

@sahilk1610

A special note, maybe for the more experienced guys in the field or also to the admin it would be great if we spend like 30 40 mins in the course to actually know how deep learning models go into production specially the ones which are using the fastai library

Yes we will be covering this.

[jeremy (Jeremy Howard)]

@oneironaut

I’ve been told that I’m good at explaining complex subjects in an easy to understand way. As such I think I can contribute the best by writing summary articles about the lectures. Probably multiple posts per lesson, given their length and information density

That would be wonderful.

@oneironaut

With that in mind: Is there any reason not to use medium? I know the paywall is frustrating. But as an author one can simply choose to not use their partner program and therefor not be listed behind the paywall, right?

Yes that’s exactly right, and it’s totally fine. Although for code and chart intensive blogging, fast_template is way faster and easier:

https://www.fast.ai/2020/01/20/blog_overview/

[jeremy (Jeremy Howard)]

@ark_aung

I want to use this course as a platform for educating students here in Yangon.

I hope you do that! I was lucky enough to visit Yangon a few years ago…

[jeremy (Jeremy Howard)]

@quan.tran

And another good news is that I am selected for USF’s Data Science Master program starting this Summer so I am extremely psyched to move to San Francisco and get to work with this year’s cohort and the fastai community there!

Oh great - I’m looking forward to meeting you! Be sure to come by and say hi.

[jeremy (Jeremy Howard)]

@AlisonDavey

I have been following on-line courses for the last couple of years and am now completing an in-person data science boot camp that I won in a Kaggle competition (thanks to fastai).

Oh wow! Please do tell us more…

[digitalspecialists (RobG)]

Hi, I’m Rob, in London UK. I’m a deep learning flaneur, leisurely strolling the streets of computer vision, now 24 months into enjoying fast.ai as a route to relax and unwind after a hard days work.

[jeremy (Jeremy Howard)]

@Brainkite

Also, I’m convinced that this field will help tackle the high dimensions problems of nowadays urbanism and city transformations that we so desperately need.

What are some ways you think that could happen?

[AlisonDavey]

I love the phrase ‘deep learning flaneur’!

[DrHB (Asimo)]

Hi Everyone!

First of all, its soo motivating and inspiring to read all of your stories. I was wondering if more people can provide there twitter (some people did) so I can follow them =)

I am structural biologist, with no prior knowledge of coding, math, deep learning and etc. I found fastai almost two years ago and immediately went thru courses and start applying on Kaggle. Last year when I took part in the course I only had 2 bronze medals and 1 silver. Since then I was able to get 1 bronze, 2 silver and 1 gold medal. I am deeply grateful to the fastai community for there encouragement, knowledge sharing and help =)

I am looking forward to take part in 2020 course =)

P.S : my twitter

[apalepu23 (Aditya Palepu)]

Thanks so much, Jeremy (and the rest of the fast.ai team and community) for this amazing opportunity! My name is Aditya, a former algorithmic trader & blockchain engineer. My passion is to leverage deep learning best practices towards socially-impactful problem areas, particularly in bias-reduction (which is also why I’m super interested in studying how bias can, somewhat ironically, creep into such models!).

I look forward to learning from the lessons and community during this newly-revamped/comprehensive course!

[shahnoza]

Hi, I am Shahnoza, I am a data analyst based in Denmark. I am a recent graduate who is building own path to Data Science career.

I started my DL journey a year ago and went through several online courses. There were no local fastai meetups in my region when I started the course, so I have decided to look for online study buddies here. Several people answered my post and now we have a beginner-friendly online meetups every week. Once, to my surprise, Jeremy tweeted about it and the study group grew big.

Thanks to this forum I got support, met amazing people, started doing DL projects and felt more confident in my abilities.

Now I am happy to be part of this highly motivated and inspiring group and looking forward to learning more with you and be an active contributor to the community.

[MicPie (Michael)]

Hello everybody,

I am super happy about the invitation for the next fast.ai course. Thank you!

My name is Michael and I am from Vienna/Austria and my DL journey started by the end of 2017 with the coursera DL course from Andrew Ng. After that I found fastai and was going through the course material several times besides learning Python from scratch.

I tried to apply my knowledge and improve my skillset in several hobby projects and kaggle competitions (some material you can find on my GitHub site). I also setup two local learning groups where we were going through the fastai and pytorch learning material (course 1 repo and course 2 repo).

I have a PhD in biotechnology but my great interest in DL helped me to get my first DL job at a very cool startup in the medical domain where I will start soon!

Thank you very much to the entire fastai community and especially to @jeremy, @sgugger & @rachel ! You had a huge positive impact on my learning journey!

I’m looking forward to the course & happy (deep) learning,
Michael

PS: Feel free to connect with me here on the forums, GitHub, Twitter, kaggle, etc.!

PPS: Because @sparalic mentioned the next ICML in Vienna/Austria: Maybe we can setup a fastai meetup for interested people in town?