About the role
Responsibilities:
- Design and implement real-time audio/video streaming systems with ultra-low latency.
- Build and optimise WebRTC pipelines, including signalling and media transport.
- Develop and maintain SFU-based architectures for scalable real-time communication.
- Ensure audio-video synchronisation and handle jitter, packet loss, and network variability.
- Configure and manage TURN/STUN servers for reliable NAT traversal.
- Develop media processing pipelines using GStreamer or FFmpeg.
- Optimise encoding pipelines using NVIDIA NVENC hardware acceleration.
- Deploy and manage streaming infrastructure on AWS (EC2 GPU, VPC, networking).
- Collaborate with cross-functional teams to integrate real-time avatar systems.
- Monitor and maintain production systems with strict latency SLAs.
Requirements:
- Strong expertise in WebRTC: ICE, DTLS, SRTP, SDP, simulcast, SVC.
- Hands-on experience with SFU frameworks: LiveKit, mediasoup, Janus, or Jitsi.
- Deep understanding of video/audio codecs: H. 264 VP9 AV1 Opus (including bitrate tuning).
- Experience with NVIDIA NVENC hardware encoding.
- Proficiency in GStreamer or FFmpeg pipeline development.
- Strong knowledge of audio/video synchronisation, jitter buffers, and packet recovery.
- Experience with TURN/STUN servers and NAT traversal.
- Hands-on experience with AWS (EC2 GPU instances, VPC, Elastic IP setup).
- Programming expertise in Go, Rust, C++, or async Python.
- Proven experience delivering sub-second latency systems in production.
Preferred Qualifications:
- Experience with CUDA-based video pipelines.
- Background in video conferencing, live streaming, or cloud gaming systems.
- Familiarity with ML inference serving: Triton, vLLM, ONNX.
- Experience with container orchestration: ECS or Kubernetes (GPU workloads).
- Contributions to open-source projects such as LiveKit, FFmpeg, mediasoup, or Pion.