• de

Making Machine Learning Applications Fast and Simple with ONNX

We are happy to start into 2024 with the next exciting event on January, 25: HPI meets... QuantCo. QuantCo will talk about "Making Machine Learning Applications Fast and Simple with ONNX ".

If you are interested in participating in the event please use our registration form. We are looking forward to seeing on January 25, at 5 pm!


Abstract of the Talk

Turning data into actionable insights is a critical requirement for many businesses. However, we often witness companies struggling to extract value due to the "last-mile problem". The integration of Machine Learning (ML) models into applications that address real-world problems on a large scale remains a complex task. It demands organizations to write ML pipelines that are highly reproducible, quick in execution, and easy to maintain. In our talk, QuantCo will show how they help organizations successfully do this by leveraging ONNX, an open-source serialization format suitable for representing ML pipelines. ONNX addresses reproducibility concerns by its ability to decouple from Python and its dependencies and also helps achieve fast execution based on compiled code. However, replicating the runtime behavior of ML algorithms via ONNX operators constitutes a significant challenge. To tackle this, they open-sourced Spox, which was designed to simplify the process of writing converters (and ONNX models in general) using clean Python. Josef Viehhauser and Chris de Vries from QuantCo will share their insights and experiences based on a real-world examples.

About QuantCo

At QuantCo, we leverage data science, engineering, and economics expertise to help organizations turn data into decisions. Started by four PhDs from Harvard and Stanford, today we are more than 170 professionals with extensive quantitative, engineering, and business experience. Our solutions include algorithmic pricing, data-driven claims management, and high-dimensional forecasting systems. Our customers include some of the largest financial, retail, and healthcare organizations in the US and Europe. We are excited about writing mission-critical code that impacts $BN's of dollars and reaches millions of people.