Viliam Tencer.developer

Software architecture and full-stack development.

Platform design, application development, systems administration and project leadership.

Application development for banking and anti-money laundering, architecture and team collaboration.

Database and application design, and SMS platform development in Java EE.

Application architecture, systems administration, virtualization and solution development.

A real-time pipeline from microphone to animated avatar: speech recognition, an LLM dialogue engine, text-to-speech and NVIDIA Audio2Face facial animation, shown on a three.js web avatar — all in Slovak.

A local web app for testing ONNX speech recognition models on files and live microphone input, comparing accuracy and speed for Slovak on CPU and GPU.

Model Context Protocol bridges that let AI agents drive professional animation and 3D software — building scenes, materials and rigs through conversation.

A staged, testable pipeline that ingests assets, assigns looks and lighting, validates the USD stage and renders it with Omniverse RTX.

A 3D side-scrolling beat 'em up and a space trading game in the spirit of 8-bit classics, with characters and assets made in Blender.

A story development studio run by AI agents — a CEO, a creative director and department roles that delegate, review and document the work.

Development of banking applications supporting anti-money laundering.

Java EE platform development, from database design to application logic.

Development work spanning cryptocurrency exchange applications and warehouse software.

Cloth simulation for character clothing on the GPU: a simulation proxy, a collision body and a cache, so sleeves, skirts and coats move with the animation in one pass per shot.

Jiggle and follow-through for hair, ears and accessories computed on the GPU and baked back to the rig.

Fixed-topology template registration: wraps a clean base mesh onto a scan or sculpt, so different characters share one topology for rigging and blend shapes.

GPU hair ribbons generated along guide curves, and generators for realistic and stylized eyebrows and tapered eyelashes.

Fixes foot sliding, floating and ground penetration by detecting contacts and planting the feet on IK controls; presets for CloudRig and Rigify.

Source-to-target bone mapping with Auto-Rig Pro mapping import and export, range baking and copy-pose-at-frame.

Visual pickers for CloudRig characters, and a tool that matches a character pose to a reference photo.

A modular foundation for humanoid pose, expression, viseme, breathing and behavior libraries, with rig mapping profiles and versioned catalogs.

AI agents build scenes, assign BlendKit materials, rig and fix characters through the Model Context Protocol, plus workspace sets that keep each task's add-ons in view.