c++ developer for conversational AI

EPAM is a global leader in digital transformation that develops enterprise software products, open source solutions, and accelerators.

• Design and develop UE5 C++ plugins and modules with runtime plugin architecture, including Build.cs configuration and module dependencies;
• Implement real-time audio streaming systems using USoundWaveProcedural, including PCM buffering, resampling, sample-rate/audio-format handling, underrun mitigation, and latency tuning;
• Integrate WebSocket and streaming protocols in C++ with persistent session management, JSON-based event protocols, reconnection strategies, and backpressure handling;
• Integrate LLM / Conversational AI voice systems, including TTS and STT, prompt and session design, turn-taking systems, and latency investigation/debugging;
• Architect event-driven systems using decoupled delegates, event bus patterns, Unreal subsystems, and cross-system integration;
• Implement MetaHuman runtime lip-sync systems, audio-to-viseme pipelines, and feed streamed audio into facial animation;
• Work with embedded Lua scripting and a Lua-driven data and logic layer, including C++/Lua bridge architectures;
• Perform audio and network debugging/profiling, including tracing raw protocol events and diagnosing audio quality, latency, and streaming issues.

• Strong experience with UE5 C++ plugin/module development, runtime non-editor code, plugin boundaries, and integration patterns;
• Proven expertise in real-time audio streaming in Unreal Engine, including USoundWaveProcedural, PCM buffering, resampling, sample-rate and audio-format handling, underrun mitigation, and latency tuning;
• Hands-on experience with WebSocket and streaming protocol integration in C++, including persistent session management, JSON-based event protocols, reconnection strategies, and backpressure handling;
• Experience with LLM / Conversational AI voice systems, including TTS and STT integration, prompt and session design, turn-taking systems, and latency debugging;
• Solid understanding of event-driven architecture, including decoupled delegates, event bus patterns, Unreal subsystems, and cross-system integration;
• Experience with MetaHuman integration, runtime lip-sync systems, audio-to-viseme pipelines, and feeding streamed audio into facial animation;
• Proficiency with embedded Lua scripting, experience working with a Lua-driven data and logic layer, and familiarity with C++/Lua bridge architectures;
• Strong audio and network debugging/profiling skills, including tracing raw protocol events and diagnosing audio quality, latency, and streaming issues;
• Nice to have: Experience with libwebsockets, experience with StateTree in Unreal Engine, background in gameplay systems and components, knowledge of multiplayer replication, proficiency with Blueprints, experience with animation systems.

• Remote work in Serbia.