Episteme99

Software-guided acceleration
of precision imaging

We are a tiny German startup developing next-generation classical and quantum software tools to accelerate precision imaging data analytics. We are based in Hamburg, Germany, and supported by an incubation grant.

To achieve scalable utilization of imaging hardware, feature-rich sensor datasets need to be processed at an optimal speed. This requirement opens up a Pareto frontier between data processing time and computational accuracy. Achieving the optimal trade-off is a critical prerequisite for real-time operando machine learning, model interpretability, and high-fidelity reconstruction. We develop specialized software tools for high-throughput precision imaging workflows, enabling users to navigate this Pareto frontier.

• Learn more

Sensor data analytics
at your fingertips

Scalable deployment and autonomous operability of imaging hardware are two major bottlenecks to delivering efficient point-of-care services. Our development is focused on providing low-latency, accurate algorithms to alleviate these challenges. We are aiming to deliver fast, reliable imaging analytics that shorten the operational cycle.
We are exploring three verticals while navigating toward product-market fit

Exploration

Advanced light sources, such as free-electron lasers (FELs) and synchrotrons, routinely map out unknown macromolecular structures at unprecedented spatial resolutions. On-demand delivery of detector data analytics can significantly accelerate the pace of identification and screening within the drug discovery pipeline.
To support this, we are developing sorting algorithms designed to reduce the number of data frames required for downstream applications.

Verification

Unaccounted errors regularly introduce production challenges within the semiconductor fabrication pipeline. Automated visual inspection of these defects can be accelerated by provisioning high-throughput, accurate analytics that work on batch datasets.
To address this, we are developing low-latency discriminator algorithms that exploit spatio-temporal features of the dataset to enable anomaly detection. The inference will lead to rapid deployment of corrective measures.

Diagnosis

The automated delivery of medical diagnostics is often hindered by inadequate featurization of microscopy datasets. Accurate real-time algorithms focusing on precise feature identification can facilitate both intuitive classification and rigorous comparison with established benchmarks.
To achieve this, we are developing specialized tools to enhance the classification of existing datasets. In future, we hope to extend these tools to include filtering and feedback control.

From workflows
to inference tools

We build customized algorithms and architectures that enable faster analytics and precise inference for high-throughput imaging. We primarily leverage high-dimensional pattern recognition, physics-informed data compression, and unsupervised learning techniques. Currently, our efforts are focused (disproportionately) on microscopy workflows.
Empowering our users to meet concurrent demands for precision and speed in imaging operations remains our singular focus.