Hi, I'm Aliaksei, a postdoctoral researcher in :elect Electrical and Computer Engineering at Western University, with a PhD background in Electrical and Computer Engineering. My work sits at the intersection of additive manufacturing, computer vision, and geometric algorithms — most recently, graph-based toolpath optimization for extrusion-based polymer 3D printing. I have hands-on expertise in pattern recognition, texture analysis, object segmentation, along with a background in sensor systems, robotics, and control. I'm currently focused on closing the loop between design, fabrication, and in-process correction — using computer vision and multi-axis robotic deposition to make polymer additive manufacturing more precise, adaptive, and material-aware.

The next step in additive manufacturing it's connecting the stages that are still solved separately today: design, toolpath planning, and in-process correction. Topology-optimized parts are still sliced as if their geometry were arbitrary; toolpaths are still generated without knowledge of how the material will actually behave; and defects are still caught after the fact, if at all. My work is aimed at closing that loop — treating geometry, material behavior, and real-time vision feedback as one connected problem rather than three separate ones. I believe the manufacturing systems of the next decade will be built this way: adaptive, material-aware, and self-correcting, from the initial design intent through to the finished part.

My research develops graph-based methods for toolpath optimization in extrusion-based polymer additive manufacturing, with a particular focus on multi-axis and robotic deposition systems. The core problem: as parts grow more geometrically complex and material systems more varied — composites, multi-material assemblies, fiber-reinforced polymers — conventional layer-by-layer slicing breaks down, and defects that emerge mid-print are rarely caught before the part is finished. I address this by representing the toolpath problem as a graph that carries geometric, material, and process constraints together, enabling paths that are optimized for mechanical performance rather than just print speed. In parallel, I develop computer vision methods for in-situ, layer-wise defect detection, aimed at closing the loop between planning and real-time correction during a print.