At the 21st WBS Quantitative Finance Conference in Palermo, our Head of AAD presented a production-ready solution for autocallable Greeks using smoothing + AAD methodology, achieving 90% computational cost reduction while maintaining accuracy for exotic derivatives pricing.

MatLogica, consistently recognized as the Category Leader in Automatic Differentiation, has also received 5 awards in total and is ranked 13th overall in the Quantitative Analytics 50 report from Chartis Research for 2025.

Independent benchmark results show MatLogica AADC outperforms popular ML frameworks such as TensorFlow, JAX and PyTorch by 40-64x for quantitative finance workloads including Monte Carlo simulations and derivatives pricing.

MatLogica was announced as "Technology Newcomer of the Year" by Asia Risk Technology Awards 2024, recognizing innovation in automatic differentiation for derivatives pricing and risk management.

This Intel technical article vividly illustrates how AADC transforms computational finance tasks by leveraging AVX512 vectorization and efficient compilation techniques with adjoint differentiation. The dramatic reduction in execution time, coupled with the ability to perform more complex derivatives pricing analyses swiftly, positions AADC as a crucial tool in the arsenal of financial institutions and quantitative analysts.

In this workshop, we provided a practical solution to address the trade-off faced by quants: the ability to prototype in Python quickly and the need for high-performing production models. We demonstrated an easy-to-use framework combining Python flexibility with AAD (Automatic Adjoint Differentiation) and "bare metal" performance, speeding up quantitative models by a thousandfold.

MatLogica is delighted to be shortlisted for the FinTech Start-up of the Year award by InformaConnect (the QuantMinds organizer), recognizing innovation in quantitative finance technology and automatic differentiation solutions.

MatLogica's Code Generation AAD™ was central to two presentations at the WBS Quantitative Finance Conference:

1. Comparing AAD Techniques & Performance - by Stephan Bosch, Quantitative Developer at ING.
2. Algorithmic Adjoint Differentiation For Tail Risk, Model Risk & Stress Testing - by Svetlana Borovkova, Probability & Partners.

Both presentations highlighted the enormous computational acceleration and efficiencies easily obtained by users of MatLogica AADC for derivatives pricing and risk management.

MatLogica receives 4 Chartis Research awards: Innovation, AAD, Data Parallel Programming, and Innovation in Computational Frameworks. MatLogica has been awarded an overall ranking of #10 out of 50 in the QuantTech rating for quantitative analytics solutions!

MatLogica graduated from the 2023 Accenture FinTech Innovation Labs programme, along with 14 other innovative FinTech companies. The program provides mentorship, industry connections, and growth opportunities for emerging financial technology leaders.

We are excited to announce that we have entered into a strategic agreement with Tachyum, a provider of unique hardware that unifies the functionality of a CPU, GPU, and TPU in a single Prodigy processor. We look forward to extending our AADC toolkit to Tachyum Prodigy, to enable native support across CPU, GPU and TPU architectures, unlocking the full potential of Tachyum's unique technology while lowering the Total Cost of Ownership for quantitative finance workloads.

MatLogica now supports AAD for CUDA! Learn how your CUDA quantitative analytics can be accelerated by AADC on CPU with an option of automatic adjoint differentiation for GPU-accelerated derivatives pricing and risk calculations.

We have produced a visual representation of how we supercharged QuantLib with MatLogica AADC for real-time risk management. Visit our homepage to interact with this live demo and talk to us to see how we can help you transform your batch risk system into live risk with intraday pricing and risk calculations!

MatLogica AADC enabled our partner, Prof. Roland Olsson, to design state-of-the-art neural network architectures for time series analysis. The results are up to 3x more accurate than available cutting-edge methods, and the training time is several times lower due to MatLogica's code generation and automatic differentiation technology.