Victor G.

I build real-time and low-latency systems: audio/video/voice, networking, synchronization, OS-level work, virtual audio devices. Co-founded Roc Streaming, open-source audio over IP: down to 10 ms ADC-to-DAC, RTP, FEC, adaptive latency. I often work with products where technical capabilities of the core define success: latency budget, real-time guarantees, system-level plumbing. I see my job as working out how far it can be pushed, and building a solution that holds it in production and can be measured. A major part of real-time work happens at design time: the architecture must keep the real-time path free of hidden locks and contention. More importantly, this should be wired into design in a way that survives project evolution. Over 15 years, I've built a toolbox of patterns and algorithms for such designs. If you want to talk about your project, feel free to drop me a line! PROJECTS: Note: The numbers below were measured on hardware+software test benches; see details in portfolio. ➤ Real-time streaming & sync (Embedded Linux, Cloud, Desktop) - Roc Toolkit: open-source real-time audio over IP (ADC-to-DAC down to 10 ms) - Device for musicians to play together online (ADC-to-DAC 30 ms) - Wireless concert wearable, network playback aligned with live sound (sync 0.25 ms) - Video feeds from cameras synchronized for remote operators (skew 0.5 ms) - Wireless audio multiroom software (sync 2 ms) - High-load trading data hub for real-time quotes (30k streams, 20k rps) ➤ OS-level & native audio + DSP pipelines (Embedded Linux, Desktop, Mobile) - Roc VAD, Roc Pulse: open-source network streaming virtual audio devices (macOS, Linux) - Roc Droid: open-source Wi-Fi audio with system capture (Android) - libASPL: open-source C++ library for virtual audio devices (macOS) - Signal Estimator: open-source latency measurement tool (Linux) - Desktop engine injecting effects into other apps' audio (Windows, macOS) - Hearing aid app for phones (Android, iOS) ➤ WebRTC & VoIP + voice pipelines (Linux, Cloud) - webrtc-cli: open-source WebRTC bridge between sound cards and browser - Voice call quality lab on real phones (WebRTC, Bluetooth) - Self-hosted speech-to-text GPU server (WebRTC, Whisper) - NAT traversal layer, direct P2P over carrier-grade NAT ➤ Embedded, RTOS, bare metal: - NTP server and client for STM32 (Zephyr RTOS) - Embedded video storage with recording and replay (Embedded Linux) - Serial over IP for industrial equipment (Embedded Linux) - Display server on USB-connected MCU (bare metal) - Millimeter-wave body scanner firmware and data acquisition (Altera SoC, uC/OS-II) SKILLS: - Real-time architectures: RT path isolation, RT-safety, arena allocators, lock-free pipelines, deadline-aware scheduling - Low-latency audio/video: adaptive jitter buffer, clock offset estimation, clock drift compensation, clock slaving, synchronization, multiroom, presentation timestamps, packet loss recovery, forward error correction (FEC), packet loss concealment (PLC) - OS-level audio: native audio APIs, high-resolution timestamps, hardware timestamps, ring buffer & DMA, sound server extensions, virtual audio devices, audio drivers, system-wide capture, loopback capture - DSP and media pipelines: resampling, mixing, surround sound, parametric equalizers, streaming statistics - Voice pipelines: voice enhancements, noise reduction, range compression, echo cancellation, AGC, speech-to-text, text-to-speech, ASR, live captioning, live transcription - Networking and protocols: asynchronous I/O, wire protocol design, RFC implementation, P2P topologies, NAT traversal - System programming: IPC, shared memory, filesystems, process control, zero-copy - Performance and optimization: multi-threading, lock-free algorithms, profiling and instrumentation, contention analysis, SIMD (NEON, SSE) - Measurements and benchmarks: latency measurement, end-to-end latency, hardware acoustic rigs, contention and throughput benchmarks, load testing STACK: Languages: C++, C, Rust, Go, Python, Java, Objective-C Streaming: AoIP, VoIP, RTP, SRTP, RTCP, ICE, STUN, TURN, SDP, RTSP, MPEG-TS, H.264, FECFRAME, Reed-Solomon, LDPC, GStreamer Audio: Core Audio, HAL Plug-In, AudioServerPlugIn, AVFoundation, WASAPI, Windows APO, ASIO, ALSA, PulseAudio, PipeWire, AAudio, Oboe, MediaProjection DSP: FFmpeg, JUCE, SoX, SpeexDSP, RNNoise, KISSFFT Voice & AI: WebRTC, Pion, libnice, libjuice, libdatachannel, libwebrtc, Opus, Whisper, Silero VAD, OpenAI, AWS Transcribe Networking: NTP, SNTP, PTP / IEEE 1588, DSCP, QoS, libevent, libuv Systems: musl, uclibc, bionic, libudev, libusb Platforms: Linux, POSIX, WinAPI, Mach, Zephyr, ARM, RISC-V, MIPS