Company:** Sibintek

Location: Moscow, Russia

Employment: full-time

Work Format: office / hybrid / remote (RF)

Our team is engaged in healthcare projects, creating breakthrough diagnostic and prognostic solutions. Our goal is to work on neurodegenerative directions: in close collaboration with doctors, we find potential opportunities and translate them into real product solutions.

Your Tasks Will Include

• Analysis and Selection of SOTA Solutions: constant monitoring of advanced approaches for specific tasks, as well as the ability to translate papers into code without ready-made repositories.

• Development and Improvement of Detection, Segmentation, and Tracking Models: training and adapting models for processing diagnostic images, video analytics, and medical history data.

• Multimodal Data Processing: development of pipelines for processing medical data from various sources.

• Implementation and Scaling: model optimization, error analysis, working with large volumes of data and training infrastructure.

We Expect You To

• Train neural network models and apply various approaches. Understand how models are structured internally.

• Work with PyTorch.

• Be knowledgeable in the field of applied DL: understand computer vision theory and be able to achieve practical results.

• Understand the specifics of real-time constraints.

• Read papers from ML conferences and follow the development of the field.

It Will Be a Plus If You

• Have worked with the medical domain.

• Have experience with large foundation models and understand how they are structured.

• Have publications at leading ML conferences.

• Hold a relevant academic degree (mathematical or physical sciences).

What We Offer

• Voluntary medical insurance with dental coverage.

• 28 calendar days of vacation.

• Annual bonus based on performance.

• Competitive salary (based on interview results).

• Corporate training and professional development opportunities.

• Access to servers with professional GPUs for training, testing, and experimenting with models.

• No restrictions on the use of modern AI tools.