Financial computation

for the future

Our mission is to revolutionize risk and regulation processes for capital markets by building streamlined and scalable solutions with integrated AI features

Product suite

Our product suite sits across three modular layers.

TensorFlow Quant Finance Library
Building blocks for banks to recreate their pricing models in TensorFlow
Pricing Engine
Efficient calculation of pricing and risk factors
Risk Reporting Engine
Regulation-aware’ engine. Currently built for FRTB-SA with plans to integrate additional regulations in the future
TensorFlow Quant Finance Library
Our open-source library provides the building blocks for banks to recreate their pricing models in TensorFlow.
Pricing Engine
These TensorFlow models, alongside portfolio data and market data, are then fed into the Pricing Engine for efficient calculation of pricing and risk factors.
Risk Reporting Engine
The sensitivities that are calculated by the pricing engine can then be used in the Risk Reporting Engine to compute aggregated metrics necessary to comply with regulations such as FRTB-SA.
Plug in + modularity

The product suite is modular and customers can plug in at each layer or use the service as an end-to-end solution.

Built-in AI features

Features such as anomaly detection and AI-driven device placement are seamlessly built in.

Synthetic portfolio & market data

The product can be used with customer data, but equally with provided synthetic portfolio & market data for ease of testing.

Features across
the product suite
  • Cloud native, cloud-hosted allowing for automatic scaling
  • Transparent data storage
  • Access to inputs and outputs via API and BigQuery
TensorFlow Quant Finance Library
Built on top of core TensorFlow
Customizable and extensible
Single code base that runs across CPU, GPU & TPUs
Computation of greeks and gradients via automatic differentiation
Pricing Engine
Automatic removal of redundant calculations for efficient pricing
Smart execution of the graph using AI
Live risk ticking graph
Risk Reporting Engine
Near real-time latency
Explanation of change in capital requirement between snapshots
Anomaly detection
Attribution of capital to specific trades or portfolios

Benefits

Scales horizontally
without impacting performance
Reduces maintenance costs
Automates operational
monitoring
Uses machine learning to increase
efficiency, improve explainability
and reduce errors
Provides incremental pricing.
You pay for what you use

Context

Stricter regulations

Recent updates to regulatory requirements such as Fundamental Review of the Trading Book (FRTB) under Basel III and Basel IV will require an estimated 5x computational power from banks and will have an impact on opex and headcount for maintenance and oversight.

Increased need for computational power

Both risk management, as a primary driver of profitability, and regulatory changes can benefit from gaining organizational and computational efficiencies.

The power of the Cloud

There is thus a pressing need for new solutions to facilitate risk calculations at an unprecedented scale while still remaining affordable - which custom value added cloud-native solutions with built in AI features can help address.

A dynamic, modular approach

Avera AI's turnkey product suite can ensure regulators no longer have to make compute-related compromises and that banks no longer need to worry about compute as a limiting factor to their business decisions.

Team

We’re a team of Googlers with finance, modelling and technology expertise partnering closely with Google Cloud. Avera AI is part of Area 120 by Google, which is a workshop for Google’s experimental products, helping small teams rapidly build new products in an entrepreneurial environment.

We’re a team of Googlers with finance, modelling and technology expertise partnering closely with Google Cloud. Avera AI is part of Area 120 by Google, which is a workshop for Google’s experimental products, helping small teams rapidly build new products in an entrepreneurial environment.

How to get started with our library
How to get started with our library

The library provides high-performance components leveraging the hardware acceleration support and automatic differentiation of TensorFlow.

Documentation
Documentation

A list of the primary symbols used in the TensorFlow Quant Finance library.

Example Notebooks
Example Notebooks

These examples demonstrate usage of the TensorFlow Quant Finance methods and can be run on Google Colab, as well as a self-study introduction to TensorFlow.

Can’t find an answer to your question?

Send us an email

Can’t find an answer to your question?

Send us an email