I’m very excited and honoured to be invited to part 2 of the 2022 course!! I’m Johannes (or Hannes) from Berlin. I am an airline pilot and made use of my suddenly plenty free time during Covid by learning Python and doing part 1 of the 2019 fast.ai course. This was a life-changing experience for me, since I realized how many opportunities there are even without a university degree.
I’m now back to flying full time, but still try to keep up at least somewhat with the latest AI trends. I’ve also started to work as a consultant, or at least I’m at the humble beginnings of it.
Hi Jeremy and co, long time fast.ai follower but first time poster (changed email and account). I am a Data Scientist from Australia. Went through the applied math route with DeepLearning.ai and Stanford, loved it. Went through the fast.ai route, also loved it. Actually ran the fast.ai course at work following the textbook and notebooks, definitely increased the output from my colleagues and even got folks involved in reading academic papers.
I’m really grateful to be here. Thank you so much to @jeremy and the Fast AI team for providing scholarships for the course.
I finished Part 1 of the course last month, learned a lot from it, and created a few projects with what I’ve learned.
I’m super excited to participate in the live course with everyone this time, looking forward to learning a ton more in the coming weeks and accomplishing great things like some of FastAI’s alumni.
Hi all! I’m Jonathan and I work in investment management in New York City. My first experience with fastai and first formal introduction to machine learning was the original Machine Learning for Coders back in 2017/2018. I’ve been following the fastai work ever since. I’m looking forward to participating in this course (but don’t think I will be around for any of the live lectures). You can find me at
Hi all. I’m also Jonathan (or Johno). I work as a data scientist and independent researcher in Zimbabwe. These days I’m quite excited about teaching generative modelling, working on my own course (very much an early work in progress) building on the previous version called AIAIART. I also blog about general data science/deep learning projects on my datasciencecastnet blog and YouTube channel.
Fast.ai was my gateway drug into deep learning, so I’m very happy to be here for this course
Apart from ML, I enjoy any excuse to get into nature - hiking, photography, rock climbing etc. In Zimbabwe nobody knows me for deep learning, I’m “that spider guy” because I do various citizen science things focused on our very under-studied insects and spiders here!
Every fastai person I’ve met has been amazing, so I’d love to meet more! Feel free to connect here or on Twitter @johnowhitaker.
I’m a data scientist with a background in molecular biology and computer vision research. 6 months ago I co-founded an AI-driven drug design company here in Sydney, Australia with VC backing - and this is now what I spend the majority of my time working on. I have to say that wouldn’t have happened if I hadn’t come across Fastai back in 2020!
This is the first time I will be venturing into part 2 of the course! Really excited to be here and dive deeper. I will always be thinking of how these techniques can be applied to biological data, if you’re in the same boat definitely get in touch!
Brilliant stuff, Elixir+OTP is one of my personal fav. languages to build networked systems. Pretty neat the the core team is putting significant efforts into bringing numerical(may i say array) programming features to the ecosystem.
Also, hats off to the the dad-son taking same live course combo.
I’m Michael, a DL research engineer in the medical / bio ML field and I recently started https://OpenBioML.org with the help of other like-minded people and stability.ai.
My ML career kind of started with Jeremy’s fastai course so I’m super happy to come back and see what has changed since my last part 2 course.
I’m looking forward to learn some new tricks from you and Jeremy.
Hey! I’m Sharwon Pius, from Kerala. I have been in and out for some time but will be focusing to be more active.
Nostalgia hitting hard, while going through the forum
This slide (lesson 9 2018 part2) had a huge impact on me. Something just clicked, and I wanted to add this skill-set to my arsenal.
Conducting a fastai meetup in Bangalore was great experience. We got exposure on how different startups were using deep learning for solving their problems. (repo)
Quick Catch-Up
I was working for Hyperverge , improving their existing model for face anti-spoofing (KYC), where the focus was to detect if a face was live or not (fraud).
Became a better DL engineer, because of fastai part2.
Worked on zero shot learning (https://arxiv.org/pdf/1904.02860.pdf), realised getting quality data for this task was quite an effort. Piqued my interest in unsupervised learning.
While implementing rPPG (extracting heart-rate from face video), came across algorithmic bias (it didn’t work well for brown faces).
Realised simple engineering tricks can help in early detection (anti-banding).
A client from a rural town used our KYC stack for sanctioning small loans; ran into a problem. A ticket was raised, but it took some time to address it. Looking at the issue; an edge of a monitor was visible in the image, which led to a false prediction (that person didn’t get a loan). This didn’t sit well with me. I didn’t know how much of an impact this had on the end customer, which led me to take a break and reevaluate.
Off-Grid Break : I was jumping ashrams in south India, and met some amazing folks.
Helped me in getting through COVID times, and also healed my sinusitis.
Last year, bootstrapped a startup along with an ex-colleague; tried to tackle the ‘visual try-on’ problem.
Started off as a fun project; helping my flatmate to expand his fashion sense.
Asked around and got inputs on how to create a business canvas (model), one minute elevator pitch … startup stuff. Surprised by the quick response that we got from Bewakoof, Myntra and Instagram Influencers.
Quickly realised that we were solving a much bigger problem. Reformulated our perspective. It became a SaaS company focused on automating the e-commerce photography industry for trendy fashion styles. We were now trying to automate a pipeline that took these companies a week to get the output. (from 5 days to 5 minutes, was the selling point)
Rest…. maybe, over a beer.
Cohesively combined 4 different models - Generative Adversarial Networks, Pose Estimation, Human Segmentation and Super Resolution Networks for validating the MVP
- noticed the super resolution network was amplifying noise from the output of the previous model.
- did a POC, found that more&better data helps.
I got into PKMs (personal knowledge management system).
After my break, felt like I was starting from scratch. I needed to have a checkpoint for my memory.
The thing which drives me; creating tools to improve human-human interaction.
I had issues with language. The medium we depend for communicating with each other. I want to find an optimal language that would help us to communicate/interact better, or contribute to that journey.
Is there a way to use all the senses for communication? Which language optimises these criteria?
Always open to a conversation regarding it.
This got me interested in robotics, specifically perception. I would like to work at the sensor level and understand how we are quantifying different input modals. It is a reflection of how we perceive the world.
I’m now working as a consultant at FloMobility , building their AI pipeline. We are trying to build an affordable autonomous navigation stack for slow-moving vehicles.
My primary focus is to build a monocular depth estimation model, which can run on a snapdragon soc.
It will eliminate the use of lidar in some applications.
Explore RGB-D SLAM and Learning based SLAM techniques.
My name is Will, I’m based in Sydney and currently completing a MSc in AI whilst working in Product Management. My undergrad degree was in econometrics and analytics. Through this I was introduced to Kaggle competitions, and it sparked a sustained interest in DL and ML.
Last year I read DL for Coders and really enjoyed the way concepts were communicated. So, when I saw this course advertised, I quickly signed up.
Looking forward to learning and collaborating with you all! If you’d like to get in touch, reach out over LinkedIn.
Hi Everyone,
My name is Harshit, I’m based in Bangalore and currently completing my MS by research at IIIT Bangalore. I got introduced to Fastai in 2021 when I was reading the notebook by Marc Auberville and his team on MIDOG 2021 challenge and have been working on mitosis detection in Brain cancer slides using FASTAI.
It is an honor to be part of the course22 and really looking forward to learning all new methods and seeing what is the hype about trendy STABLE DIFFUSION models.
Thanks, everyone for this wonderful community.
Hi everyone,
I’m Luke. I’m currently skilling up to work in AI Safety. I come from a math & philosophy background and work in software. I run a weekly discussion group about AI Safety - let me know if anyone wants to join.
Hi everyone, Very excited to be added to this course. I am a resident physician (PGY-3) working in Radiation Oncology. I was introduced to FastAI in 2018. I previously used FastAI to develop models for automatic diagnosis and segmentation of prostate cancer in prostate MRI as well as image quality improvement and super resolution. In that process I developed my own integration of DICOM data into the fast.ai library which has fortunately been replaced by Jeremy’s much better integration recently.
I’m looking forward to this course and hoping to use what I learn to be able to predict tissue changes following radiation treatments particularly in brain tumors.
Hi, I am Slava.
I work in finance and very excited about Deep and Reinforcement Learning. I went through Fast AI in 2018 first time and cannot stop myself experimenting with image recognition and deep learning since then. That changed my life and my career. I made my path from lawyer to software engineer/researcher. I do highly appreciate what Jeremy and his team does, proving that this subject is for everyone who passionate enough about it!
I am goutham. I was introduced to fastai in 2017. Been a huge fan of their teaching style.
Looking forward to participate and learn great stuff in coming weeks.
I’m Prajwal, a Software Engineer from Bangalore, India.(LinkedIn , GitHub)
I have been following previous offerings of the course and loved the top-down approach of teaching. I’m super excited to participate in the live course
