My name is Jean de Dieu Nyandwi. I am a graduate student at Carnegie Mellon Africa, in MS Engineering AI. I am broadly interested in deep learning, computer vision, and multi-modal learning(most notably vision-language).
Happy to be here and excited for the upcoming course! I have heard many good & practical things about fast.ai and I look forward to learning from everyone in this amazing community!
I am Aakash Nain. I have been working in the field of DS/ML for almost 5+ years. I am one of the maintainers of TensorFlow-addons package, contribute to Keras and JAX. Here is a list of some of my works that you might be aware of:
In the past, I have collaborated with @Ekami@jamesrequa@Sayak, and @init_27 I am always open to discussions related to Machine Learning, Deep Learning, MLOps, API designs, etc. Here is my Twitter handle in case you want updates related to latest research.
Fast.ai has been and always is a constant source of inspiration for me and all of us that wish for AI to be used to make a better world, and it is always a great pleasure to be involved in whatever fast.ai organizes,
I am Javier Rodriguez and right now I head the ML team at a start up called Assetario.
My degree is in theoretical physics and I moved to the private sector a few years ago after a decade in academia (to which I am still tangentially related via a research associate position at the Center for Astrobiology in Spain )
I am also the creator and main contributor of the library pytorch-widedeep. This is a library intended to facilitate the combination of tabular, images and text datasets, and is in constant (and active) development.
I did Fastai v1 and v2 courses (when they publicly released it). Then I scanned through the following versions always trying to watch videos, or run notebooks whenever I had a sec. The truth is that, even if I was familiar with the concepts, I always learned something that I could bring to my library or to work. There are plenty of jewels hidden in the library (e.g. that Tokenizer!), or a fast, clever implementation of something that Jeremy came up with.
Pretty excited to go through one of the courses again.
Hello to the greatest online learning community! It’s so exciting and an honor to be invited back to what will surely be yet another mindblowing fastai course
For those who don’t know me, I started my DL journey 6 years ago, joined fastai 5 years ago, and basically I owe everything I know about ML/DL/AI to Jeremy & Rachel and the fastai community! Its quite surreal to see all of the familiar faces here after all these years. Probably the most rewarding part for me is witnessing how many people came into this community completely new to DL who have now blossomed to become the top experts in the field and leaders of this community (@init_27@radek@sgugger@ilovescience just to name a few!!)
For the past several years, I have been working full-time as an AI engineer focused in the area of cancer detection. Its been really special to be able to work alongside fellow fastai alumn like @Ekami and collaborate with many others from the fastai community on Kaggle comps & side projects (@alexandrecc@nain@sermakarevich)
Hi all! I am so excited to participate in this course again! Like many who’ve commented before me, I have benefited tremendously from the Fastai course and community since I first joined in 2018. It’s no stretch to say that my experiences here have been life-changing: I am now a tenure-track astronomer at the Space Telescope Science Institute and an associate research scientist at Johns Hopkins University. My research focuses on understanding how galaxies evolve by using a combination of astronomical observations and deep learning.
In previous threads I’ve said a bit more about my specific research projects, but maybe I’ll add another shameless plug… Earlier this year my collaborators and I used convolutional neural networks (a la Fastai) to distinguish rare, low-mass, nearby galaxies from the overwhelming number of faint, distant, background galaxies (paper here). The figure below shows our method in action: left panel shows examples from the training sample (~100k) and the right panel shows examples the test sample (>4M); upper panels show CNN-classified distant galaxies, while the lower panels shows nearby dwarf galaxies. This method allows us to study the relative distributions of “satellite” dwarf galaxies around more massive galaxies in a more comprehensive way than ever before!
Thanks again to Jeremy and the many others who have enabled a new (and growing) cohort of expert machine learning practioners! Feel free to get in touch with me on Twitter or LinkedIn if you’d like.
Hi all! I’m Johan von Forstner, currently working as a MLE at Paradox Cat, a software supplier in the automotive industry, in Munich, Germany.
My academic background is in Space Physics, where I analyzed data from energetic particle detectors on several different space missions, most recently ESA’s Solar Orbiter. In early 2021, I left academia and started my deep learning journey, joining Paradox Cat to build cool prototypes of in-cabin sensing systems for a large German automaker.
The fast.ai Part 1 course materials have been extremely helpful to start off with this field, and as I am also using the fastai library at work I have been able to contribute back a couple of improvements and bugfixes. Here at Paradox Cat, I’ll soon be organizing an internal study group with my colleagues for the new fast.ai Part 1 course.
In my spare time, I often contribute to open source as well, and as an electric vehicle nerd I’m the maintainer of a charging station finder app for Android (which does not contain any AI so far though ).
I’m super excited to be able to follow the Part 2 course live now! Thanks a lot to Jeremy and the team.
Awesome to see so many familiar faces here! I’m a fast.ai alum from the very first series, and deep learning with fast.ai changed my life.
My area of research is recommender systems, and I lead a team at NVIDIA writing open source software for training and deploying recsys called Merlin. It’s written with similar principles to the fast.ai library in the sense that we’re trying to build something that is both easy to use for beginners, and possible to modify deeply for experts. Like fast.ai it’s open source and we welcome contributions. One of our most popular libraries is a dataloader for tabular data for TF and PyTorch that’s 10x faster than native dataloading for training. @radek is one of the amazing people on the team and his contributions to the fast.ai community were one of the reasons I pursued him.
I’m Erick Muzart, an auditor and data scientist in the Brazilian Court of Accounts, in Brasília.
I helped to found the Deep Learning Study Group in Brasília that was the biggest group of fast.ai students in South America. We’ve followed most of fast.ai courses editions, helping hundreds of colleagues go through all this great content.
I am looking forward for this newest edition of DL part 2 to reach the state of the art with Transformers, LLM and diffusion models!
I’m Andy Rowlinson. I’m excited to be joining this course and have benefited so much from the previous versions.
I live in Finland . I am active in the football/soccer analytics community and have made mplsoccer for data visualization. I also helped setup the soccermatics course for working with football data. You can follow me on twitter @numberstorm.
Hi all. Thank you, Jeremy once again, it’s a pleasure to have the opportunity to join the course. I see some familiar faces around, glad to see you guys.
I am Sergii. Fastai DL course helped me to move from tabular DS to CV and later I moved from CV to ML engineering. Currently I am working on fares optimization platform at Flyr.
I’ve been working closely with @jeremy and @hamelsmu on nbdev and other fast.ai related projects for the last few months. Before that I worked in the South African AI startup scene . It’s little known, but in a past life I played computer games professionally
I’m super excited to be joining my first live fast.ai course! Especially since we’ll be digging deep into the foundations together
I’m most active on twitter (@wasimlorgat) and less frequently write (wasimlorgat.com). I love chatting about software and AI, so please feel free to get in touch!
Hello! I’m absolutely ecstatic to have the opportunity to participate in this course!
I’m Ali and I live in Maryland. My journey into ML and Software in general has been driven in large part by the energy and ideas permeating this community. I am a Computer Engineer at the Berkeley Lab — where I have mainly worked in genomics. I also am interested in the ethics surrounding ML.
In my free time I love reading, yoga and spending time with friends + family.
I’m @azaidi06 on twitter — where I do need to get myself more active
*This is what first piqued my interest into the field. It was only after Jeremy convinced me – that I could apply it myself – that I decided to give it a try
I’m Chris, and I work as a Machine Learning Engineer/Scientist as part of the Microsoft Data & AI Service Line in London, UK.
I have followed every Fastai course since the Keras days back in 2017, all of which have played a huge part in moving into my current role, so I am eternally grateful to Jeremy and the team for putting out such awesome content. This will be my first time attending one of the courses ‘live’, so I’m excited to get involved as we are going along, rather than a couple of months later!
My ML interests are predominantly in Computer Vision and Recommender systems, I maintain a lightweight PyTorch training library (based on HuggingFace accelerate) which is used by myself and colleagues in cases where Fastai doesn’t quite fit, and write blog posts occasionally.
Hi Diego, @pdillis
Happy to meet with you somewhere in the middle. Hopefully we find other people and make a larger group. I am happy to present some of the background papers that Jeremy posted if that is useful.
I’m Ben Mainye, a microbiologist and data scientist from Kenya . I have been in a couple of iterations of the course in the past. I am working with people from a Research Institute in Kenya called Institute of Primate Research regarding making diagnostic tools that can be used in hospital and in research; We use fastai a lot too in the work. Here’s one tool we are working on called a Cell explorer. I have been trying to come up with effective data science infrastructure to support the project above and slowly getting back to doing bioinformatics/genomic data science to enrich the project. As a start we are looking into antibiotic resistance using familiar methods that are talked about in the course & book too with a couple of students I am mentoring.
Thanks for inviting me back to part 2 of the course @jeremy and the community. Looking forward to engaging and learning. Stable diffusion is
(Had to delete my previous post as I think I replied personally to Jeremy Sorry, still getting used to the interface!)
Hey all
I’m Misa Ogura, an MLE at Healx where I use graph ML for drug discovery - so excited to be here!
A bit more about myself - I’m originally from Tokyo and came to the UK for undergraduate studies in Biochemistry, followed by postgraduate studies in Cancer Cell Biology. Like with many fellow fast.ai folk, I come from a non-traditional background without a degree in CS or ML and made my way into a career with ML with self-studying, scholarships and side projects such as FlathTorch.
This will be my first fast.ai course - looking forward to getting my teeth in!
Hi everyone! Miko (Michelangelo) here. Italian mathematician turned product manager turned data scientist in London. I’m on my second run on the live fast.ai course, but this is the first part 2 that I get into. Which to me is EXTREMELY exciting as part 2 of 2019 (last one that has been made) was one of the most interesting courses I have ever watched and really got me into thinking better about Deep Learning.
I have recently gotten into privacy and AI, and how to train/deploy on confidential data using secure enclaves.
For this purpose, we have released BlindAI, an open-source solution to deploy AI models in two lines of code with privacy guarantees for the data owners sending data to the AI model. We can see it illustrated below with an example of what happens with/without BlindAI for speech analysis in the Cloud.
We have also silently released BastionAI, a Federated learning framework using secure enclaves, inspired by fastai. We have an example notebook on how to fine-tune a BERT model on a private dataset with Differential Privacy.
I am happy to be back in the course, learn the cutting edge AI and help the community in my own means.
Hi! I’m Charlie. I’m an engineering manager at Cruise working on self-driving cars (specifically on the Machine Learning Infrastructure team). Prior to Cruise I worked at an ML-ops startup called FloydHub, and plenty of other places too. I love to write, read, hack on little projects, and I’m excited about this course! Here’s my blog: charlieharrington.com and I’m @whatrocks on Twitter. Excited to learn with everyone!
).
I’m Andy Rowlinson. I’m excited to be joining this course and have benefited so much from the previous versions.
. I am active in the football/soccer analytics community and have made 
. It’s little known, but in a past life I played computer games professionally 
Especially since we’ll be digging deep into the foundations together
. I have been in a couple of iterations of the course in the past. I am working with people from a Research Institute in Kenya called Institute of Primate Research regarding making diagnostic tools that can be used in hospital and in research; We use fastai a lot too in the work. Here’s one tool we are working on called a 
