Senior Full-Stack & Systems Engineer — real-time media, streaming & graphics
Founder of Nixosoft · shipping production software for 19+ years
I build production systems end to end — from GPU-accelerated media pipelines and real-time streaming to a from-scratch motion-graphics engine, the distributed infrastructure that runs it all, and the web apps on top. I like hard problems at the boundary of performance, graphics, and systems, and I ship them: live products with real users, not demos.
- 🎥 Real-time media & streaming — WebRTC, HLS/LL-HLS, FFmpeg, NVENC/NVDEC, adaptive bitrate, OBS-style compositing
- 🎨 Graphics engines — a from-scratch Rust compositor/renderer (CPU + GPU path tracing via
wgpu) targeting After Effects–class capability - 🖥️ Distributed & remote systems — browser-delivered desktops, multi-server orchestration, self-hosted platforms
- 🌐 Full-stack web — React / TypeScript / Node, SSE-driven live dashboards, auth & SSO, clean APIs
- ⚙️ DevOps — Docker, nginx, TLS, CI/CD, GPU hosts, monitoring, zero-drama deploys
🟢 Currently open to senior / staff engineering roles in media, streaming, graphics, or systems.
⚡ 8,500+ contributions in the last year — I ship, daily.
I design and ship AI-powered products, not just prototypes — LLM agents, retrieval pipelines, speech systems, and generative interfaces running in production, with the infrastructure to serve them reliably.
- 🧠 LLM apps & agents — RAG pipelines, tool-using agents, and conversational assistants on modern LLMs (LangChain / LangServe, hosted & local models), deployed serverless on AWS Lambda
- 🗣️ Speech & audio AI — text-to-speech services and LLM-powered subtitle & transcription generation for live and offline media
- ✨ Generative & applied AI — generative UIs, AI-driven job matching, and AI features embedded directly into shipping products
- ⚙️ Production ML delivery — self-hosted inference on GPU hosts (Ollama / local model serving), clean APIs, and the DevOps to keep it up
| Project | What it is | Stack |
|---|---|---|
| Heddle | A from-scratch motion-graphics & compositing engine — masks, effects, 3D/2.5D, a built-in path tracer (BVH + GI, GPU via wgpu), and its own scene language. Aims at After Effects / Blender–class rendering. | Rust, wgpu, WASM |
| Nixosoft Platform | A self-hosted application platform & web desktop — SSO, containerized apps, full Linux desktops in the browser. | Rust, Docker, nginx |
| WebStation | Low-latency browser-delivered remote desktop over WebRTC with adaptive FEC & encoder bitrate control and a live Chromium UI. | C++, GStreamer, WebRTC |
| CustomHLSStreamer | Production Node.js service for customizable live HLS streams with real-time content updates. | Node.js, FFmpeg, ZeroMQ |
| media-stream-orchestrator | A headless, OBS-like library: multiple scenes, layered sources, and streaming — no UI. | Node.js, FFmpeg |
| Nixosoft | My software studio — the umbrella for the products above. | — |
Much of my recent systems and engine work ships through private repos; the links above are the public-facing products. Happy to walk through architecture on request.
Let's build something.
hiradnikoo.com ·
nixosoft.com ·
LinkedIn ·
@HiradNikoo ·
YouTube