Senior Systems Engineer - C++ / Real-Time Audio

Vaanis orchestration stack is roughly 148,000 lines of C++ - written deliberately, before agentic coding tools existed, because sub-400ms end-to-end latency on real phone lines is not something you get from glue code. You will own parts of that engine.

This is the layer that makes the difference between a voice agent that feels human and one that people talk over. Every millisecond you claw back is visible in production metrics the same week.

+
• Own latency end to end - profile, and remove, the microseconds across the audio pipeline from carrier ingress to speech egress
+
• Build and harden real-time audio paths: jitter buffers, barge-in and interruption handling, streaming inference, backpressure
+
• Work on hybrid inference - deciding what runs where, across cloud, VPC and on-prem deployments
+
• Keep the engine deployable inside customer infrastructure, including air-gapped environments
+

+
• Deep modern C++ (17/20) and genuine comfort with concurrency, lock-free structures, and memory behaviour under load
+
• Experience with real-time media, telephony, or streaming systems - WebRTC, SIP, RTP, or equivalent
+
• You profile before you optimise, and you can prove the improvement
+
• Bonus: GPU inference serving, ONNX/TensorRT, or Linux performance tuning at the kernel boundary
+