The Atom team provides updates on Hydrogen’s core API suite, as well as financial engineering and general architecture initiatives.
Welcome to the first official update from the Atom team!
Atom forms the core of the Hydrogen API suite, and it is used by developers around the world to build and deploy financial services applications. Atom offers the tools and the foundation to create digital platform across Investing, Savings, Financial Wellness, Insurance, Financial Planning, and more.
There are two main components within Atom: Nucleus and Proton.
Nucleus is used to manage data relating to fundamental parts of a financial application. This includes clients, accounts, financial products, on-boarding questionnaires, money transfers, and a variety of other entities.
Proton houses business logic and computations that developers can use to power their products and applications. This includes things like simulations, risk assessment models, and optimization frameworks. Please see our developer portal for more details on Atom.
As this is the very first update from the Nucleus team, we want to give readers a little bit of background. Earlier this year, we worked tirelessly to release a v1 of the Nucleus API, which provides a core digital infrastructure to anchor fintech applications.
Nucleus v1 is already being used by major financial institutions around the world to build new applications for their customers, and we’re currently preparing a more robust developer program to give startups and individual developers the ability to access the API in a pay-as-you-go model. Most recently, over the past week, our team has been working on critical items as we move toward a version 1.1 release.
The primary focus has been on developing endpoints for an “orchestration” API layer to encompass some key workflows required for applications built with Atom. Some examples of these workflows are On-Boarding, Portfolio Management, and Goals-Based Wealth Management. By coordinating data across multiple underlying services, orchestration can help simplify common operational processes.
The team has also been working on enhancing endpoints used to calculate financial metrics such as rate-of-return, volatility, alpha, beta, and Sharpe ratio. Our goal is to make these services more streamlined and easier to use, and a big part of that is adding the ability to retrieve multiple statistics with a single request.
We’re gearing up for the release of version 1.1 of the Proton API! The team has been hard at work finalizing development, testing, and documentation for the new 1.1 services.
The new release is set to include a collection of financial calculators. With these tools, developers will be able to compute a variety of things for their users, including some of these examples:
- Emergency fund savings schedules
- Amortization metrics for mortgages
- Breakdowns of effective life insurance needs
- Projections on savings accounts
- Plans to achieve goals (like retiring or making major purchases)
This is our first release following the initial launch of the Proton API, so we’ve been refining our development cycle. We want to make sure it goes without a hitch. Proton has two main pieces — there’s the Algorithm layer (which handles the underlying logic and computations), and the API layer (which handles user requests/responses and communicates with data resources as needed). Because we have different criteria and testing frameworks for each layer, it’s important that we follow a rigorous QA process in which both layers are vetted.
Over the past few weeks, the team has implemented some performance enhancements on the API’s Goals endpoints. We have modified some of the underlying algorithms — which are implemented in Python — to compile down into C-compatible code using Cython. As a compiled language, C is optimized for lower-level operations and can execute significantly faster than Python for certain processes.
In preparation for future expansions of the Proton API into more insurance-related services, we have also started working on a framework to handle actuarial calculations in the Algorithm layer — with the help of our newest team member, Dan. This framework will be used to derive things like biometrics, insurance pricing, and annuity outcomes.
Look for Proton v1.1 to be rolled out in the coming days, and stay tuned for many more exciting things coming to Nucleus!