A short post to explain why I even put my book reviews on here
Posts by Year
Quick reference on a generalized network setup based on number of layers
Linear Algebra reference especially useful when vectorizing Machine Learning operations and understanding intuition for algorithms
Backprop can be error-prone. Numeric methods for calculating the gradient is computational slow but easy to implement
Knowing which activation functions are available, and how it’s changed over the years, gives us choices for optimization and inspiration for new functions
Using the following metrics gives a more accurate picture of model performance
Marketing and selling high-tech products to mainstream customers.
Understanding the activation of a network with the help of semantic dictionaries and non-neg matrix factorization to increase network performance.
An approach to help leaders recognize and attain leadership balance required for teams
How can we make sure the representation learned by our model will be applicable to data the model has never seen before. Overfitting can reduce the validatio...
A set of strategies for weight initialization preventing DNNs from having problems like vanishing or exploding gradients
As an engineering team leader, this book gives great advice. Some of it is outdated, but much of the relevant topics are included below
Gradient Class Activation Mapping provides a method that can be used to visually explore possible biases in the CNNs and DNNs
Simon explores what it means to become a truly admired leader
Knowing what hyperparameters are available for tuning is an important part of quickly and accurate optimization
Understanding the spatially important features of Deep and Convolutional Neural Networks using Class Activation Mapping
Extracting visual insight from network layers allows us to optimize and investigate our models
The rules that should govern the actions of an engineering team
A review and summary of one of the foundational startup books.
Good organization is key when it comes to the performance of our engineering team. When it comes to training and validation of an ML algorithm, there are som...
Defines the leadership style of the 90’s and what can be learned and used in the 20th century
Positive leaders can transform teams and organizations
Picking the right ML algorithm requires careful consideration of the user, product, team, and business
Understanding how to iterate through the values of the hyperparameters when tuning a Machine Learning algorithm can make a difference when it comes quickly t...
Washington ‘presented’ as the first founder and CEO of two startups; the army and the presidency.
Biomedical engineers go through a multidisciplinary training that might make them a good major to transition into machine learning and Artificial Intelligence
An examination of Lincoln’s leadership abilities and how they can be applied to today.
As will any software engineering field, it’s important to consider technical tradeoffs so that we can maximize customer value.
Focusing on the module with the most potential gives direction on where to optimize the model pipeline
Listed below are values that are important in leading a successful engineering team when the goal is to create a useful product.
Process of developing and delivering backlog stories as a cohesive and efficient engineering team
Process of how to create a high value prioritized backlog of understandable stories
A quick book review from the perspective of the management of a engineering team.
Explore why great companies can do everything right and then still lose their market
Most innovations in the digital age were done collaboratively. This is the story of why their ability to work as a team made them so creative and successful
Discussing importance in terms of other features vs in terms of an arbitrary scale keeps focus on the criterion used when creating the feature vs the arbitra...
An explanation of the two systems that drive the mind
Focusing on how you think is more important in productivity than what you think
Learning by examples
Quick little cheatsheet to get you through ES6
Most startup failures are preventable. How to leverage both human creativity and capital more efficiently
A review and summary of one of the foundational startup books.
Access the MongoDB instance on your MUP (Meteor UP) deployment using a GUI explorer
Client side generation of PDFs using jsPDF, HTML, DataURIs, and Cordova on iOS8+ and Android in MeteorJS
Quick examples of using bluebird with mongoose in node. I just want to show you a couple common use cases to help you recognize how to use promises.
When your @IBDesignable is either too complex or requires data not available when previewing in the interface builder, here is how you deal with those errors.
An easier way to remember what each vim command does
It turns out that there is enough difference in the properties each framework uses that it can be frustrating to learn how to use the ios-charts library. Her...
Two design patterns which can be used to implement Singletons in Swift.
How to create and class a 100% customized NSNumberFormatter in Swift. Useful when frameworks (cough ‘ios-charts’) require a NSNumberFormatter for value forma...
A common use case is to use Dokku without an actual domain. I address how to use a consistent external port with an ip address since the docs don’t.
UIScrollView is something that can be very confusing because it uses constraints in a way that they aren’t normally used. Because of this there is so much co...
Quickly deploy MeteorJS to your own provider or services like Digital Ocean.
Attributed strings build in the storyboard editor which use a custom font does not render correctly. Here are the steps to get around it
Promises seem like an amazing way to take advantage of non-blocking io in a very readable way
A little information on how to edit your swagger.json document (v2) so that you can accept file uploads in NodeJS. Includes a workaround for a current bug.
How various iOS apps implement SMS/Text based sharing. We consider the pros and cons of each.
Building a Chrome extension that changes more than just the font of a webpage allowing dyslexic users to have a better experience regardless of page design.
Awful routing makes me wonder if I am being gouged by Uber in cities I am not familiar with.
An alternative to npm libraries like ‘jasmine-where’ and full blown data provider pattern. Anything that gives us a boost for our TDD & CI/CD pipeline is...
f you aren’t with it the following with work with built in vim :substitute
An alternate config file in node for setting variables based on current env without the overhead of another library with more flexibility than a .json file.
The syntax is a little different since ‘type’ already has a meaning in a Mongoose schema.
A quick code example of how to embed a subdocument/collection of objects in a Mongoose model without creating a whole other explicit subdoc reference
Discussing a bug as a feature in the Lyft app that potentially gets users rides at a lower ‘prime time’ multiplier.
Continuous integration and test-driven development are important strategies for any engineering team that wants to be efficient.
Forget about finding a deeply nested id property that you have to call the .toString() method on in order to compare.