AI Voice Model Consultant based on the real experience of Voice AI & Audio Synthesis

We are a forward-looking team building the next generation of creative AI audio tools. Our mission is to create personalized, high-fidelity AI voice experiences that empower artistic expression. We are seeking a seasoned AI Voice Model Expert to architect and build a groundbreaking system for AI-generated singing and voice interaction.

The Core Project:
Your primary mission will be to design and develop a system that allows a user to sing any song in their own voice. This involves capturing a user's unique vocal tone, timbre, and singing style from recordings and applying it to generate high-quality vocal covers for any chosen song.

Key Responsibilities:
You will be responsible for the end-to-end development lifecycle, from research and prototyping to deployment and MLOps.

1. Voice Cloning & Singing Voice Conversion (SVC/RVC):
-   Research, adapt, and implement state-of-the-art models for voice cloning and singing voice conversion (e.g., RVC, So-VITS-SVC, DiffSinger, DDiFFSinger, VoiceLab).
-   Build robust pipelines for dataset preprocessing, feature extraction (e.g., F0, hubert features), and model training on user-provided voice data.
-   Focus on achieving exceptional quality, capturing nuances like vocal inflection, vibrato, and emotional delivery.

2. Text-to-Speech & Speech-to-Text (TTS/STT):
-   Integrate or develop TTS components for voice agent functionalities, ensuring natural prosody and alignment with the user's cloned voice characteristics.
-   Utilize STT systems for potential lyric alignment, transcription, or interactive voice commands.

3. Voice Agent & Interactive Systems:
-   Design the architecture for a voice agent that can handle user requests (e.g., "generate a cover of Song X," "make the voice sound more powerful").
-   Create a seamless workflow for song input, lyric synchronization, and vocal generation.

4. Music & Lyrics AI:
-   Explore and integrate models for music source separation (e.g., Demucs) to isolate instrumental backing tracks and original vocals.
-   Investigate AI for lyrical analysis, alignment, and potentially even lyric generation or style transfer to match a user's style.

5. MLOps & Engineering Excellence:
-   Architect, build, and maintain scalable ML pipelines for training, fine-tuning, and inference.
-   Implement model versioning, monitoring, and automated retraining pipelines.
-   Containerize models (Docker) and deploy them on scalable cloud infrastructure (e.g., AWS, GCP, Azure).
-   Ensure the entire system is reliable, efficient, and maintainable.

Deliverables & Outcomes:
-   A scalable, cloud-based platform for user voice model training and inference.
-   A well-documented API and/or a simple, user-friendly interface for end-users to generate new vocals.
-   A library of trained, high-quality voice models, starting with the founder's voice.
-   A robust MLOps foundation for continuous improvement and scaling of the AI capabilities.

Must-Have:
    -   5+ years of experience in ML engineering with a strong focus on generative AI and deep learning.
    -   Proven expertise in digital signal processing (DSP) and audio/music processing.
    -   Hands-on experience with SVC/RVC frameworks and a deep understanding of the underlying architectures (e.g., GANs, VAEs, Diffusion Models).
    -   Strong proficiency in Python and core ML libraries (PyTorch, TensorFlow).
    -   Solid experience with MLOps tools (e.g., MLflow, Kubeflow, Weights & Biases) and cloud deployment.
    -   A strong portfolio or examples of past projects in AI voice synthesis, music generation, or a closely related field.
Highly Desirable:
    -   Experience with TTS/STT systems (e.g., Tacotron, WaveNet, Whisper, VALL-E).
    -   Knowledge of music information retrieval (MIR) and lyrics processing.
    -   Experience building interactive voice/AI agents.
    -   Understanding of ethical AI principles, especially concerning voice cloning and deepfakes.

How to Apply:
Please submit the following:
1.  Your resume/CV.
2.  A link to your portfolio, GitHub, or demos of relevant work (e.g., AI-generated vocals, SVC models, deployed ML systems).
3.  A brief cover letter describing your specific approach to building a custom singing voice conversion system, including the tools and models you would prioritize and why.
4.  [Optional] Any open-source contributions or research papers in related fields.