Technical Workshops

In this workshop, we will cover starting, experimenting, serving and scaling a machine learning solution. We will take a representative problem and cover the following phases of machine learning:
Data Analysis/Inspection: We will perform data analysis (cover few key strategies of data analysis prior to machine learning)
Feature engineering: We will uncover a few techniques of feature engineering, learning of the previous step and approaches for feature engineering that will help you improve data quality prior to kicking off machine learning.
Representative model selection: In this phase we will summarize model selection criteria and select a couple of models suitable to our problem.
Training, Serving and Scaling in Cloud: In this phase we will uncover model evaluation metrics, understand serving skew and ability to scale both training and serving of Machine learning solutions.

In this workshop, we’ll be covering how Fitbit Development SDK works and what tools we offer as part of our complete developer environment. We’ll focus on how the SDK allows us to research and market health solutions and we’ll deep dive with a case study of one of the health apps that we offer to our user

The workshop will be hosted by engineers that are working on the Fitbit Developer Platform (FDP) team. This team is responsible for the development and maintenance of the Fitbit Public SDK (on device JavaScript Runtime & Developer Environment). The SDK is used for developing 1st, 2nd and 3rd party apps for our smartwatches. This is a cross functional team, with members working on mobile, firmware, backend and frontend.

The current landscape contains a large variety of architectures ranging from monoliths, to service-oriented, to microservices to serverless architectures. When creating a new fintech product, a developer needs to figure out a solution design that will fit all requirements. At the same time we need to check all the boxes for characteristics like: future-proof, stability, security, cost, easy CI/CD integration, etc. Decades worth of challenges pushed systems architects to seek advantages over all these different solutions and we’re going to take a look at two of the currently most discussed solutions.