Deep Learning on AWS - Tips and Tricks

This is my list of hints and tips for this course. It’s markdown so you can save it, access it or store it anywhere. I might also give you other links that are course specific. I’ll add specific answers to questions I get during the course. I’ll share it with everyone.

Your Instructor

Ian Falconer https://www.linkedin.com/in/leftbrainstuff/

Administrivia

We need to jump through some hoops to get access to the labs, notes and my hints and tips. Be consistent with the email address you use for all sites. There are three seperate sites you need to access and one bitly link which is this page:

Join or login to https://www.aws.training/ to ensure your training and certifications are captured. No we don’t spam you or sell your details.
Access Qwiklab (yes it is spelt INCORRECTLY)
- aws.qwiklabs.com for the labs in this class
- run.qwiklabs.com for outside of the class or to do other labs at your own pace. NOTE: Some are free others require course credits. Also check out the AWS Professional Developer Series of MOOCs on edX https://www.edx.org/aws-developer-professional-series
Access the course notes and slides. You’ll receive two emails. One confirming your attendance at this course and with the following links. The download link seems broken. You can download apps for phones, tablets and laptops. Or use your browser.
www.vitalsource.com look for a signup link and download link. Or just go to https://evantage.gilmoreglobal.com/#/user/signin
Once you’ve logged into Vitalsource (aka Bookshelf, Gilmore, eVantage) you can redeem your unique course materials code (in a seperate email) and update your book list. You should see a lab guide and student guide for Deep Learning on AWS, version 1.4 . The student guide is the powerpoint decks and notes and the lab guide is the step by step instructions for the labs. The lab guide is included in the labs so this document is somewhat redundant. You can download the Vitalsource Bookshelf app for Windows, Mac, IoS and Android at https://support.vitalsource.com/hc/en-us/articles/201344733-Bookshelf-Download-Page
You can print the student and lab guides to pdf from the app.

Deep Learning Resources

Dive into Deep Learning, An interactive deep learning book with code, math, and discussions that is authored and maintained by Amazon AI specialists. https://d2l.ai/. Think textbook and Jupyter notebooks integrated in a learning portal.
MIT Online Deep Learning book / labs https://www.deeplearningbook.org/ This is a good read for anyone wanting a refresher or first broad look at Deep Learning.
CNNs Explained in a relatively short blog post interspesed with graphics. https://www.freecodecamp.org/news/an-intuitive-guide-to-convolutional-neural-networks-260c2de0a050/
- Note that CNNs can parellize, within a sequennce, but cannot capture sequential information like a long sentence
- RNNs however can capture sequential information but aren’t able to parallelize with a sequence A transformer combines CNN and RNN for NLP. Well suited to translation and difficult transcription workloads
Andrew Ng’s Convolutional Neural Networks Course on Coursera. https://www.coursera.org/learn/convolutional-neural-networks/home/welcome He has a ton of experience and his explanations on Deep Learning are easy to comprehend

Academic papers

Many of the key Artificial Intelligence (AI), Machine Learning (ML) and Deep Learning (DL) algorithms and theory papers are decades old. Here’s a selection of papers for those who want to dive deep.
Machine Learning, Neural and Statistical Classification (circa 1994) edited by Michie, Spiegelhalter and Taylor http://www1.maths.leeds.ac.uk/~charles/statlog/whole.pdf . Includes detailed comparisons between classification algorithms on a range of datasets available in 1994. Structured as a Journal of academic papers.
Information Theory, Inference, and Learning Algorithms (7th Ed 2005) By David McKay. http://www.inference.org.uk/itprnn/book.pdf . This book has been revised every year since 1995 (to 2005 as of this printing) It’s a solid treatise on information theory fundamentals and proofs with some theoretical discussion of neural networks and sparse graphs.
Understanding Machine Learning: From Theory to Algorithms http://www.cs.huji.ac.il/~shais/UnderstandingMachineLearning/
Decent summary of scene recognition identification challenges. Ground Truth Data, Content, Metrics, and Analysis https://www.embedded-vision.com/sites/default/files/apress/computervisionmetrics/chapter7/9781430259299_Ch07.pdf
Lab academic papers:
- Lab 3 Deep Residual Learning for Image Recognition https://arxiv.org/pdf/1512.03385.pdf

Data Sets

Finding suitable training data sets can be a huge challenge. AWS provides access to many large data sets
- Latest public data sets from AWS https://aws.amazon.com/about-aws/whats-new/2019/01/aws-public-datasets-now-available/
- Registry of Open Data on AWS https://registry.opendata.aws/
- AWS Awesome Day public data sets titled Awesome Public Datasets https://github.com/awesomedata/awesome-public-datasets and the git clone link https://github.com/awesomedata/awesome-public-datasets.git
The 50 Best Public Datasets for Machine Learning https://medium.com/datadriveninvestor/the-50-best-public-datasets-for-machine-learning-d80e9f030279
List of Public Data Sources Fit for Machine Learning (BigML) https://blog.bigml.com/list-of-public-data-sources-fit-for-machine-learning/

Cool links

AWS has now released free digital ML training (more than 45 hours) and announced the new Machine Learning Certification https://aws.amazon.com/blogs/machine-learning/amazons-own-machine-learning-university-now-available-to-all-developers/ and https://aws.amazon.com/training/learning-paths/machine-learning/
AWS created solutions, reference architectures and quickstarts at https://aws.amazon.com/big-data/getting-started/tutorials/ . Look for the self paced labs that are accessible at run.qwiklabs.com
There are many interesting solutions on the AWS Machine Learning blog at https://aws.amazon.com/blogs/machine-learning/
Access hundreds of ML algorithms and models from the AWS Marketplae https://aws.amazon.com/about-aws/whats-new/2018/11/awsmarketplace-makes-it-easier-to-build-machine-learning-applications-on-amazonsagemaker/
Which services are available in which regions? https://aws.amazon.com/about-aws/global-infrastructure/regional-product-services/ ReInvent is always a time for ML announcements:
- For ReInvent 2019 checkout https://aws.amazon.com/new/reinvent/ -ReInvent 2018 was a huge event for ML. 13 major ML annoucements by AWS. https://www.businesswire.com/news/home/20181128005660/en/Amazon-Web-Services-Announces-13-New-Machine and https://aws.amazon.com/new/reinvent/?sc_icampaign=event_reinvent2018-recap&sc_ichannel=ha&sc_icontent=awssm-1690-default-hero&sc_iplace=hero&trk=ha_awssm-1690-default-hero
8 Machine Learning Algorithms explained in Human language (sort of) https://www.datakeen.co/en/8-machine-learning-algorithms-explained-in-human-language/
AWS provided ML resources on Github. https://github.com/aws-samples/machine-learning-samples which includes useful tools and the comprehensive Amazon Sagemaker collection including algorithms https://github.com/awslabs/amazon-sagemaker-examples
Transcribing podcasts is a good example of serverless architecture. It would not take much effort to make this production ready. It can be really efficient to move from Development to Production grade solutions. Another benefit of serverless you may not have considered. https://aws.amazon.com/blogs/machine-learning/discovering-and-indexing-podcast-episodes-using-amazon-transcribe-and-amazon-comprehend/
The Neural Network Zoo depicts many interesting and useful neural networks Images from http://www.asimovinstitute.org/neural-network-zoo/
Build your own real-time voice translator application with AWS services https://aws.amazon.com/blogs/machine-learning/build-your-own-real-time-voice-translator-application-with-aws-services/. This blog post has a Cloudformation launch button which will build this solution for you.
Code example for setting up an adaptive learning rate in Python Using Learning Rate Schedules for Deep Learning Models in Python with Keras https://machinelearningmastery.com/using-learning-rate-schedules-deep-learning-models-python-keras/
Machine learning was used in a Japanese Decoding Dreams experiment. Interesting read https://www.the-scientist.com/notebook/decoding-dreams-39990
Deep Dynamics Models are an interesting use of deep learning methods to help control movement of exoskeletons, juggling balls and other movement tasks. Structured Deep Dynamics Models For Robot Manipulation http://rse-lab.cs.washington.edu/se3-structured-deep-ctrl/ Check out the video of a robot poking objects
Speech and Speech Pattern recognition latest research. Check out Table 1 which compares approaches and amounts of data needed. The paper is titled DEEP DISCRIMINATIVE AND GENERATIVE MODELSFOR PATTERN RECOGNITION https://pdfs.semanticscholar.org/cea9/c5f7117b3db7e62f35b4d290cfb84ddd7ba3.pdf

Visualizing a Deep Learning Model at work

Using style images (say a Van Gogh or Salvador Dali image) to morph another image to that artistic style in an article titled Artificial Intelligence meets Art: Neural Transfer Style https://towardsdatascience.com/artificial-intelligence-meets-art-neural-transfer-style-50e1c07aa7f7 . Here’s the underlying academic paper which visually depicts how the deep learning network is learning the style. https://arxiv.org/pdf/1508.06576.pdf . This paper contains image matrices that let us see how the network learns.

Controversy around AI (Dealing with the Hype)

Does AI Truly Learn And Why We Need to Stop Overhyping Deep Learning https://www.forbes.com/sites/kalevleetaru/2018/12/15/does-ai-truly-learn-and-why-we-need-to-stop-overhyping-deep-learning/#7d9ca7a68c02
Interoperability - Open Neural Network Exchange https://onnx.ai/ and https://github.com/onnx/onnx
Choose your ML frameworks carefully. One size does not fit all.
- Tensorflow sucks is a blog post that highlights when I wouldn’t use tensorflow and some of the benefits of other frameworks http://nicodjimenez.github.io/2017/10/08/tensorflow.html
- What Java needs for true Machine / Deep Learning support https://medium.com/@hsheil/what-java-needs-for-true-machine-deep-learning-support-1571ffdbb594

Governance of Data Science

The Data to Decisions Cooperative Research Centre published a Data Science Competency Framework framework available at https://uploads-ssl.webflow.com/5cd23e823ab9b1f01f815a54/5d0076903a1e4f6bb45ea50b_Data%20Science%20Competency%20Framework.pdf and here is the entry link https://www.d2dcrc.com.au/article-content/data-science-competency-framework

Comparisons of ML and DL

Amazon Sagemaker Developer Guide is a great read to start with. Both on the service and algorithm specifics. Here’s an example of the Linear Learner Algorithm https://docs.aws.amazon.com/sagemaker/latest/dg/linear-learner.html The dev guide explains the algorithm, it’s use and compute selection. This is a very useful guide for anyone looking to run ML models. https://docs.aws.amazon.com/sagemaker/latest/dg/sagemaker-dg.pdf The Amazon Sagemaker Developer Guide also includes detailed information on permissions and security of Endpoints, Notebooks and AWS services used by Sagemaker.
Comparing Deep Learning Frameworks. Contains most of the frameworks but is limited in terms of helping you choose algorithms to solve specific problems. https://en.wikipedia.org/wiki/Comparison_of_deep_learning_software
Choosing Algorithms
- Sagemaker’s built in algorithms have concise examples of usage and brief descriptions. https://docs.aws.amazon.com/sagemaker/latest/dg/algos.html
Check out Kaggle to see competitions and applications for ML and DL techniques. https://www.kaggle.com/competitions
Difference Between Softmax Function and Sigmoid Function and when would I use each. http://dataaspirant.com/2017/03/07/difference-between-softmax-function-and-sigmoid-function/

ML and DL Libraries, Frameworks and APIs

ML and DL Libraries
- XGBoost - for structured and tabular data. Popular in Kaggle competitions.
- TensorFlow - Symbolic math library for dataflow programming. Implements the define-and-run or static-graph approach. More of a black box approach.
- PyTorch - Combines Python and Lua for performance. For computer vision and neural linguistic programming. Supports the define and run approach and provides more visibility into how the model learns. Uber’s Pyro probabilistic programming language software uses PyTorch as a backend. Caffe2 has been merged into Pytorch as of Mar 2018.
Deep Learning Frameworks
- Caffe - Vision, Speech and Multimedia. Caffe can process over 60M images per day with a single NVIDIA K40 GPU. That’s 1 ms/image for inference and 4 ms/image for learning and more recent library versions and hardware are faster still. Caffe is often considered to be among the fastest convnet implementations available.
- MXNet - For CNNs and LSTM deep learning. Cloud support and by AWS. Here’s a good primer on MXNet https://mxnet.incubator.apache.org/versions/master/faq/why_mxnet.html
- Chainer - Python specific including Numpy and Cupy. First to implement define-by-run or dynamic-graph approach, the connection in a network is not determined when the training is started. The network is determined during the training as the actual calculation is performed. Works well with the Nvidia Cuda GPU library
Optimizing Compilers
- Theano - Python library and optimizing compiler for manipulating and evaluating mathematical expressions, especially matrix-valued ones. Especially symobolic differentiation. Implements the define-and-run or static-graph approach
ML and DL APIs
- Keras Keras is a high-level neural networks API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano. It was developed with a focus on enabling fast experimentation.
- CUDA - NVIDIA C++ GPU max bandwidth toolkit https://devblogs.nvidia.com/even-easier-introduction-cuda/
- Gluon is an imperative API for MXNet that’s flexible and easy-to-use

ML and DL Architectures on AWS

Near real time analysis of streaming video https://aws.amazon.com/blogs/machine-learning/analyze-live-video-at-scale-in-real-time-using-amazon-kinesis-video-streams-and-amazon-sagemaker/ This option seems to simplify all previous solutions. The KIT is new as of 19Nov2018 and retrieve the Cloudformation templates from https://docs.aws.amazon.com/kinesisvideostreams/latest/dg/examples-sagemaker.html#examples-sagemaker-create_ (https://aws.amazon.com/blogs/machine-learning/analyze-live-video-at-scale-in-real-time-using-amazon-kinesis-video-streams-and-amazon-sagemaker/)
TensorFlow deployment to AWS Deep Lens https://aws.amazon.com/blogs/machine-learning/deploy-your-own-tensorflow-object-detection-model-to-aws-deeplens/
Combining Human input with Supervised Learning on Amazon Sagemaker https://aws.amazon.com/blogs/machine-learning/use-amazon-mechanical-turk-with-amazon-sagemaker-for-supervised-learning/
Analyze Amazon Neptune Graphs using Amazon SageMaker Jupyter Notebooks https://aws.amazon.com/blogs/database/analyze-amazon-neptune-graphs-using-amazon-sagemaker-jupyter-notebooks/
Serverless Sagemaker Architectures
- Serverless SageMaker Training and Deployment Orchestration https://github.com/aws-samples/serverless-sagemaker-orchestration
- Load test and optimize an Amazon SageMaker endpoint using automatic scaling https://aws.amazon.com/blogs/machine-learning/load-test-and-optimize-an-amazon-sagemaker-endpoint-using-automatic-scaling/
Predictive Maintenance entry page on AWS https://aws.amazon.com/mp/scenarios/bi/predictive/
Using AWS IoT for Predictive Maintenance a blog post from AWS https://aws.amazon.com/blogs/iot/using-aws-iot-for-predictive-maintenance/ and check out the links in the Learn more section including https://aws.amazon.com/iot/solutions/industrial-iot/
Machine Learning Techniques for Predictive Maintenance lists some methods and specific maintenance use cases. https://www.infoq.com/articles/machine-learning-techniques-predictive-maintenance/
Using Graph Databases with ML and DL models:
- Here is the link to the AWS graph database examples on Github https://github.com/aws-samples/amazon-neptune-samples
Financial use of ML and DL
- How Patreon Avoids Fraud While Funding the Emerging Creative Class https://aws.amazon.com/blogs/startups/how-patreon-avoids-fraud-while-funding-the-emerging-creative-class/
- FICO: Fraud Detection and Anti-Money Laundering with AWS Lambda and AWS Step Functions https://aws.amazon.com/blogs/architecture/fico-fraud-detection-and-anti-money-laundering-with-aws-lambda-and-aws-step-functions/
- Virtual credit cards in Expanding Credit Card Horizons with Extend https://aws.amazon.com/blogs/startups/extend-virtual-credit-card/
- A novel way to acquire ground truth on credit card fraud in Fraugster Separates Credit-Card Fraudsters From Mere Frequent Fliers https://aws.amazon.com/blogs/startups/fraugster-separates-credit-card-fraudsters-from-mere-frequent-fliers/
- Identity verification in emerging markets in Aella Credit empowers underbanked individuals by using Amazon Rekognition for identity verification https://aws.amazon.com/blogs/machine-learning/aella-credit-empowers-underbanked-individuals-by-using-amazon-rekognition-for-identity-verification/
- Ensemble Learning — Bagging, Boosting, Stacking and Cascading Classifiers in Machine Learning using SKLEARN and MLEXTEND libraries. https://medium.com/@saugata.paul1010/ensemble-learning-bagging-boosting-stacking-and-cascading-classifiers-in-machine-learning-9c66cb271674 which talks
- Also search across AWS Blogs, AWS Marketplace, Case Studies, Quickstarts, Solutions, Solution Space and the Compliance pages using the financial keywords.
  - Some useful case studies include https://aws.amazon.com/solutionspace/financial-services/
  - Security & Compliance for Financial Services and dive deeper for handbooks and auditor tools and attestations https://aws.amazon.com/financial-services/security-compliance/

MLOps

The ML, DL and AI space is not well supported with mature developer tools for managing infrastructure, CICD, delivery and deployments. An emerging capability around MLOps is gaining traction. But ML CICD and Ops is fundamentally different to what devs and applications know and do today.
From AWS re:Invent 2019: Continuous deployment for ML: The new software development lifecycle (DEM30-S)
- Machine learning (ML) will fundamentally change the way we build and maintain applications. How can we adapt our infrastructure, operations, staffing, and training to meet the challenges of the new software development lifecycle (SDLC) without throwing away everything that already works? ML is the future of application development, but presently, many ML teams are flailing without a process or trying to shoehorn their ML workflow into tools that don’t fit the requirements.
- https://www.youtube.com/watch?v=_ImmIRDHUtY&list=PLWaYLZud5zZlsPSM1IWfbXzEwqBtgcx_Z&index=38&t=0s

Object Detection

Build your own object classification model in SageMaker and import it to DeepLens https://aws.amazon.com/blogs/machine-learning/build-your-own-object-classification-model-in-sagemaker-and-import-it-to-deeplens/
Identifying bird species on the edge using the Amazon SageMaker built-in Object Detection algorithm and AWS DeepLens https://aws.amazon.com/blogs/machine-learning/identifying-bird-species-on-the-edge-using-the-amazon-sagemaker-built-in-object-detection-algorithm-and-aws-deeplens/
Deploy your own TensorFlow object detection model to AWS DeepLens https://aws.amazon.com/blogs/machine-learning/deploy-your-own-tensorflow-object-detection-model-to-aws-deeplens/
Build your own object classification model in SageMaker and import it to DeepLens https://aws.amazon.com/blogs/machine-learning/build-your-own-object-classification-model-in-sagemaker-and-import-it-to-deeplens/
Here is a mix of AWS Solutions and github resources for serverless image handling
- facial_recognition using Tensorflow and CNN https://github.com/ruby913/facial_recognition
- Serverless Reference Architecture: Image Recognition and Processing Backend https://github.com/aws-samples/lambda-refarch-imagerecognition

Time Series Forecasting

Kinect Energy uses Amazon SageMaker to Forecast energy prices with Machine Learning https://aws.amazon.com/blogs/machine-learning/kinect-energy-uses-amazon-sagemaker-to-forecast-energy-prices-with-machine-learning/
Amazon Forecast is a fully managed service for time series forecasting using a range of prebuilt algorithms. See Choosing an Amazon Forecast Algorithm for more details https://docs.aws.amazon.com/forecast/latest/dg/aws-forecast-choosing-recipes.html
Forecasting financial time series with dynamic deep learning on AWS https://aws.amazon.com/blogs/machine-learning/forecasting-time-series-with-dynamic-deep-learning-on-aws/
Analyze data in Amazon DynamoDB using Amazon SageMaker for real-time prediction https://aws.amazon.com/blogs/big-data/analyze-data-in-amazon-dynamodb-using-amazon-sagemaker-for-real-time-prediction/
Some LSTM models for time series data:
- A chainer sentiment analysis notebook https://github.com/awslabs/amazon-sagemaker-examples/tree/master/sagemaker-python-sdk/chainer_sentiment_analysis
- Now available in Amazon SageMaker: DeepAR algorithm for more accurate time series forecasting https://aws.amazon.com/blogs/machine-learning/now-available-in-amazon-sagemaker-deepar-algorithm-for-more-accurate-time-series-forecasting/

Cost Management of AI

Reducing deep learning inference cost with MXNet and Amazon Elastic Inference https://aws.amazon.com/blogs/machine-learning/reducing-deep-learning-inference-cost-with-mxnet-and-amazon-elastic-inference/

Building ML / DL Containers on AWS

Building your own algorithm container https://github.com/awslabs/amazon-sagemaker-examples/tree/master/advanced_functionality/scikit_bring_your_own/container

Sagemaker and Jupyter Notebooks Links

Spark with Jupyter Notebook inside a VPC https://bytes.babbel.com/en/articles/2017-07-04-spark-with-jupyter-inside-vpc.html
Jupyter Notebooks are Breathtakingly Featureless — Use Jupyter Lab https://towardsdatascience.com/jupyter-notebooks-are-breathtakingly-featureless-use-jupyter-lab-be858a67b59d This blog post highlights the use of Notebooks for exploratory data science and how JupyterLabs is more suited for building production solutions.

Compute links

Here’s a useful sortable table of EC2 instance types, sizes and specifications. Not your add columns like available for EMR which makes instance selection very simple. https://ec2instances.info/

Graph Database Links

Using Amazon Sagemaker to host a Jupyter notebook for loading data, storing graph queries and plotting of charts. No ML in use but another use of Sagemaker to handle the undiffereniated heavy lifting of your infrastructure so you can focus on just running graph queries. https://aws.amazon.com/blogs/database/let-me-graph-that-for-you-part-1-air-routes/

Programming Constructs

Python NDArray https://docs.scipy.org/doc/numpy/reference/arrays.ndarray.html

Self paced Learning and Building

You can try out Jupyter notebooks online at https://jupyter.org/try
AWS Certification roadmap https://aws.amazon.com/certification/ Check out the learning paths link at the bottom of the page.
Read the service FAQ pages, http://aws.amazon.com/faqs/, and documentation for each of the services. Just search for AWS + name + documentation in any search engine. You can keep the documentation as pdf, html online or even in your Kindle. You can also git clone the documentation for most services.
Find and build interesting AWS and partner solutions you find the in AWS Blog https://aws.amazon.com/blogs/ . Any post you find with a yellow launch button will build that solution using Cloudformation.
AWS free digital training is mostly 100 level but we also have over 40 hours of Machine Learning training available for free. You can search by topic, role or level. https://www.aws.training/LearningLibrary?src=courses You’ll find specialist deep dives from level 100 through 300 like this video describing the differences between NACLs and Security groups. https://www.aws.training/Details/Video?id=16486 NOTE: You’ll need to enroll and allow popups in your browser.
You can also take AWS Qwiklabs Labs for free at https://aws.amazon.com/training/self-paced-labs/
Get a sandbox or personal account. There are free tiers for many services. https://aws.amazon.com/free/
http://run.qwiklabs.com and complete quests and labs. These enhance your familiarity with AWS services without you having to use your own account. Some labs are free. Others will require you to redeem Qwiklab credits. Reach out to your training manager or AWS account manager. Also check out the Exam guides for SA, SysOps and Advanced Networking https://www.amazon.com/Certified-Advanced-Networking-Official-Study/dp/1119439833/ref=sr_1_1?s=books&ie=UTF8&qid=1519925473&sr=1-1&keywords=advanced+networking
Search github, https://github.com/aws , and the AWS blogs, https://aws.amazon.com/blogs/ , for solutions that interest you. Look for posts with a launch button. These will build a complete environment using Cloudformation. Retrieve the Cloudformation templates either from the built environment in your account or from Github. You can reverse engineer or use these templates as scaffolds for your own use.
Visit Stackoverflow and the AWS discussion forum to pose questions or to contribute to answers about AWS
You can also take a number of AWS MOOCs (Massive Open Online Courses) on EDx and Coursera including:
- EDx https://www.edx.org/school/aws
- Coursera https://www.coursera.org/aws
There are many other self paced labs and solutions you can build on AWS. Try:
- Build a Serverless Web Application https://aws.amazon.com/getting-started/projects/build-serverless-web-app-lambda-apigateway-s3-dynamodb-cognito/
- How about AWS Developer Center https://aws.amazon.com/developer/ where you can build the Mythical Misfits app in your choice of programming language.
The AWS Podcast has a monthly update which is a great way to keep up with the latest changes, releases and interviews with domain experts https://aws.amazon.com/podcasts/aws-podcast/
AWS has released a number of webinars and now has a monthly cadence https://aws.amazon.com/about-aws/events/monthlywebinarseries/
AWS Answers is now available to the public. It contains some interesting links. https://aws.amazon.com/answers/
Get to know your AWS Solution Architects and your Technical Account Manager (TAM). The SAs help you to architect and understand best practice. The TAMs provide support for your applications running on AWS. They can help you prepare for major events like testing and scaling. They can also help troubleshoot and provide visibility into AWS infrastructure metrics for troubleshooting. https://aws.amazon.com/premiumsupport/faqs/
AWS Glossary contains service names and nomenclature https://docs.aws.amazon.com/general/latest/gr/glos-chap.html
Now go build stuff…

Continue reading articles in my Amazon Web Services series