Vladimir B.

15+ years shipping production systems. I build AI for business first: agents, RAG, orchestration, and C#/.NET inference that has to stay up after release. When the product needs real 3D, I also work in Unity (games, XR, in-product 3D). If it needs a web or backend layer, I own that too. What I deliver - LLM apps & agents: tool calling, multi-step workflows, planner/executor loops, human-in-the-loop review - Orchestration: Semantic Kernel, graph-style agent flows, custom runtimes, queue/worker pipelines (Temporal / Hangfire / BullMQ) - RAG: chunking, embeddings, hybrid search, reranking, grounded answers over docs, CMS, or game data - C# / .NET AI: Semantic Kernel, Microsoft.Extensions.AI, OpenAI & Azure OpenAI, ONNX Runtime, ML.NET - Unity AI: Unity Sentis / Inference Engine, on-device & server-side inference, AI NPC/dialogue, editor & content copilots - Realtime AI: streaming chat, STT/TTS voice flows, WebRTC + LLM backends for live experiences - Production layer: auth, rate limits, cost controls, tracing/evals, fallback models, safe prompt & tool boundaries Unity, 3D & XR Unity LTS · C# · URP/HDRP · Addressables · Timeline · Cinemachine · Shader Graph · custom tooling · CI build pipelines OpenXR · XR Interaction Toolkit · AR Foundation · Meta Quest · HoloLens 2 · Apple Vision Pro Netcode for GameObjects · Photon Fusion/PUN · Mirror · PlayFab · WebSockets · WebRTC Around the experience (when needed) React / Next.js · Node.js/TypeScript · PostgreSQL · Redis · Docker · AWS · CI/CD · Sentry / Grafana Selected results - Mixed-reality museum installation in under 10 weeks - thousands of daily visitors, no downtime - Internal Unity toolchain that cut level-design iteration time significantly - Legacy PHP monolith to Kubernetes services with much faster, safer deployments - WebRTC multiplayer backend scaled to high concurrency on AWS spot capacity How I work I use modern AI tooling to move faster, but I still own architecture, performance, tests, and release quality. Small iterations, early measurement, maintainable code your team can keep after handoff. Best fit Business AI (agents, RAG, internal tools) as the main job. Unity/3D when the product needs a real scene, not just a dashboard. Or both, under a hard deadline. If you share the goal, constraints, and timeline, I can send a short scoping note on what’s feasible and how I’d approach it.