It is back and it is in person, it is re:invent! Now comes the biggest challenge, working out where to go and which sessions to catch. Therefore I have put this re:invent session guide/recommendations list together so that if your are into ML Ops, ML Architecture, Edge Computing or Data Analytics then it can act as a starting point. However please leave sometime in your schedule because as AWS announce new stuff they will add more sessions.
Hang on, what if you are not going, then your are like me. I have to stay home for some personal reasons. However you can still catch the keynotes and leadership sessions on the live stream. Do not forget the big three keynotes; Adam Selipsky (taking over from Andy), Werner Vogels (always dev centric), Swami Sivasubramanian (all ML). Later you can also use this guide to watch the Breakout Session once AWS have made them available on demand.
|AIM320||Implementing MLOps practices with Amazon SageMaker||Breakout Session||Invite|
|AIM407||Train deep learning models at scale with Amazon SageMaker||Breakout Session||Invite|
|AIM405||Right-sizing Amazon SageMaker compute instances for ML model training and inference||Chalk Talk||Invite|
|AIM413||Detect machine learning model drift in production||Chalk Talk||Invite|
|DOP211||Building scalable machine learning pipelines||Chalk Talk|
|ARC323||Designing Well-Architected machine learning workloads||Chalk Talk|
|KYN003||Swami Sivasubramanian Keynote||Keynote|
|AIM401||Create, automate, and manage end-to-end ML workflows at scale||Workshop||Invite|
|IOT306||Building a people counter with anomaly detection using AWS IoT and ML||Workshop|
Out of these 10, my highlights and must see list, that I can not wait to catch is:
- Implementing MLOps practices with Amazon SageMaker : ML Ops is not all about pipelines, however they are one of the key technology enablers. This session is looking at Amazon SageMaker pipelines and how to rapidly deploy your models
- Detect machine learning model drift in production : One of my personal passions I speak about is the 360 view of a model. This is the business impact, the ML performance (accuracy, etc), System Performance, Inference Observability, and Cost. This session will look in detail at ML performance and detecting drift
- Designing Well-Architected machine learning workloads : AWS has just released an update to the Well Architected ML Len and this session will cover all the best practice about how to lower your costs, keep your security rock tight, cope with failures, scale to millions of predictions and keep a watchful eye.
So what about the parties and the fun! If your in Vegas checkout:
- Midnight Madness
- Amazon’s World Famous Chicken Wing Eating Contest
If you are staying home:
- Checkout your local user group. I know some are doing local watch parties! thats a nice way to still feel part of it.
- Another fun thing also I know is happening is ComSum (Community Summit UK) www.comsum.co.uk will be back in action. ComSum will be giving you a byte size digest of the days big news and expert punditry on everything happening in Vegas. This will include myself returning as one of their ML and Data Pundits. More details soon!
Lastly, Don't worry I will be at re:invent next year, so watch out Vegas!