AI Engineer

The project is for a foreign consumer tech client developing a mobile application for real-time, two-way speech translation over cellular networks. The engineering challenge focuses on achieving extremely low latency (target sub-1.5 seconds time-to-first-audio) under real network conditions (LTE/5G). The architecture involves streaming pipelines: STT (Speech-to-Text) → MT/LLM → TTS (Text-to-Speech) orchestrated in Python and transmitted via WebRTC.

• Documented commercial experience building production real-time streaming audio pipelines; batch or offline STT insufficient.
• Practical knowledge of at least two streaming components (such as STT, MT, TTS) and orchestration tools (e.g., Pipecat, LiveKit, Vapi, or custom Python layers).
• Production experience with WebRTC, including handling audio transport under degraded network conditions.
• Experience measuring latency on real mobile devices, analyzing percentile distributions (p95), and tail behavior.
• Experience selecting AI providers compliant with strict data protection norms (EU hosting, zero data retention).
• Fluent English (minimum C1) for technical discussions and documentation.
• Clear and visible real-time audio/voice pipeline experience on CV, subject to client technical approval.

Nice to have:

• Knowledge of React Native and native audio modules (Swift, Kotlin).
• Experience with Bluetooth audio on iOS/Android.
• Experience with on-device AI inference or voice cloning (speaker similarity).

• Manage and verify per-component latency budgets (VAD, transport, STT to first token, MT, TTS to first audio, jitter buffer, playout).
• Evaluate providers for STT, MT, and TTS respecting streaming interfaces and compliance requirements (EU data residency, zero data retention).
• Design automated testing framework for AI-to-AI call evaluation and define quality gate criteria based on p95 measurements on physical devices.
• Prepare foundational codebase including repository skeleton, CI/CD pipelines, and initial architecture for Alpha build deployment.
• Collaborate directly with the client's Fractional CTO on architectural discussions and defend technological decisions.
• Produce clear architectural specifications and reports in English.

The data controller is co.brick Sp. z o.o. located in Gliwice, Poland. Applicants have rights concerning their personal data as per GDPR, and data will be processed for recruitment purposes only. Providing data required by Polish labor law is mandatory; refusal may hinder recruitment. Personal data will be stored during recruitment and for possible claims period; consent for future recruitments can be withdrawn anytime.