Data Science UA is a service company with deep expertise in AI and Data Science. Our story started in 2016 with the first Data Science UA Conference in Kyiv, and since then, we’ve built one of the largest AI communities in Europe.

The company is on a mission to make the world’s industrial environments safer with technology solutions and to drive a paradigm shift in safety from reactive to proactive approaches to reduce the risk of industrial accidents. The team in Ukraine is developing the future of industrial safety based on machine learning, computer vision, and IoT.

We’re looking for a systems-minde Middle Software Engineer to build the performance-critical core of our real-time computer vision pipeline. You’ll own the runtime beneath the CV engines – shared-memory frame pools, zero-copy IPC, and process orchestration – where the bottleneck is memory bandwidth and microseconds, not network or nodes.

– Collaborate with CV/ML engineers to turn model inference code into reliable production pipeline stages.
– Develop and optimize the core Pipeline Engine – an in-process, real-time runtime that runs the CV pipeline as a graph of stages, where latency is measured in milliseconds.
– Use and extend the zero-copy shared-memory pools that pass frames between processes (C core + Python/Go bindings).
– Build and tune low-latency ZeroMQ communication layers for passing control messages, metadata, and synchronization signals between pipeline components.
– Work with pipeline metadata formats and serialization boundaries, with awareness of performance tradeoffs around FlatBuffers or similar schema-based binary formats.

Must-have:

– Middle-level engineering experience in performance-sensitive systems, with strong systems thinking: concurrency, process models, memory ownership, data movement, latency/throughput tradeoffs, failure modes, and profiling. Practical Python experience is expected, but the core requirement is systems-level reasoning.
– Experience building or operating real-time ML/CV inference pipelines, video processing systems, robotics perception stacks, or other latency-sensitive data-processing runtimes.
– Solid experience with Docker and containerized environments;
– Solid Linux fundamentals.

Nice-to-have:

– Basic understanding of CUDA, including IPC handle management, stream synchronization;
– Knowledge of C/C++/Go.
– Experience with video/CV infrastructure: OpenCV, GStreamer, FFmpeg, RTSP streams, or camera pipelines.

The client offers:

– Good compensation;
– Benefit package;
– Strong team and career growth;
– Challenges every day!