Milosz Dudek

SoftServe Middle R&D Engineer

Milosz Dudek is an R&D engineer specializing in speech and audio AI, multimodal machine learning, and Generative AI. His work focuses on building end-to-end AI systems that combine speech processing, large language models, computer vision, and sensor data to solve real-world business and research problems. His recent work includes speech recognition and diarization, voice-based intelligent assistants, multimodal sensing, explainable AI, and efficient deployment of deep learning models across edge and cloud environments.

He also works on evaluating and adapting foundation models for multilingual and domain-specific applications, as well as designing low-latency AI pipelines for interactive systems. Milosz combines research-oriented experimentation with practical engineering, focusing on turning state-of-the-art machine learning methods into scalable prototypes and production-oriented solutions.

Milosz’s recent stories

By Marian Petruk, Milosz Dudek, Alesandro SlepcevicOct 02, 2026

Build Fast, Reliable Agentic AI From Edge To Cloud

SoftServe and Akamai benchmarked governed AI agents at 22ms and $0.0047 per 1,000 images. See what we found.

Don't want to miss a thing?

Milosz’s recent stories

Build Fast, Reliable Agentic AI From Edge To Cloud