Machine Learning Engineer (Voice AI & Audio)

#vacancy #ml #nlp #machinelearning #mlengineer #voiceai #tts #python #audio #remote #itjobs

➕ About the project: We are an innovative product company developing cutting-edge generative audio and voice transformation technologies. We do not rely on formal labels (Junior/Mid/Senior) as they are often misleading indicators of actual skill. Instead, we are looking for a candidate based on their real, hands-on confidence in voice synthesis and Python, regardless of their official "grade."

💼 Responsibilities:

— Develop, train, and fine-tune machine learning models for voice synthesis (TTS) and AI voice conversion — Design, build, and maintain robust ML pipelines from data preparation to production
— Optimize audio processing models for streaming and real-time inference to achieve ultra-low latency
— Turn research-grade models into reliable production web services and REST APIs
— Analyze and improve the quality of synthesized audio, ensuring natural sound and high fidelity
— Collaborate with the team to seamlessly integrate ML models into the final user-facing product

🧠 Requirements:

— Strong, foundational knowledge of Python — this is an absolute prerequisite no matter what your level is
— Proven hands-on experience and strong skills in Voice Synthesis (TTS)
— Deep understanding of AI Voice Conversion algorithms and audio signal processing
— Experience building and deploying complete ML-pipelines
— Practical experience with streaming optimization and real-time audio ML serving
— Hands-on experience with modern deep learning frameworks (PyTorch, NeMo, HuggingFace, etc.)
— Ability to write clean, production-ready code (moving beyond Jupyter Notebooks)
— English — B2+

💫 Nice to have:

— Experience working with voice/audio models as part of a broader, more complex ML stack
— Background in Natural Language Processing (NLP) or Automatic Speech Recognition (ASR)
— Experience with backend frameworks (e.g., FastAPI) and containerization (Docker)
— Familiarity with audio libraries like PyDub, Librosa, or similar
— Experience with MLOps practices and tracking tools (Weights & Biases, etc.)

CVs to @vladiskashh