Member of Technical Staff – Post‑Training Applied Audio

About the role

Liquid‑AI is scaling its end‑to‑end multimodal speech and text model LFM2.5‑Audio (1.5 B parameters) for enterprise customers. This position focuses on post‑training work that adapts the model for real‑time, on‑device deployment, turning spoken user intents into structured tool calls.

Key responsibilities

• Own the full lifecycle of audio post‑training projects, from requirement gathering to delivery and evaluation.
• Translate customer use‑cases into concrete specifications and workflows for LFM2.5‑Audio and future models.
• Design and implement function‑calling capabilities that map spoken intents to API invocations, parameter extraction, and confirmation flows.
• Build data‑generation pipelines for speech‑to‑speech and text‑to‑text training, including synthetic dialogues and intent‑action pairs.
• Run supervised fine‑tuning, preference alignment, and reinforcement‑learning pipelines on audio language models.

Required profile

• Demonstrated ownership of end‑to‑end ML projects, especially in audio or speech domains.
• Ability to reason across data pipelines, speech‑text alignment, model adaptation, and evaluation as an integrated system.
• Pragmatic mindset focused on model quality and customer outcomes over academic publications.
• Comfort with on‑device constraints such as low latency, limited memory, and privacy requirements.

Required skills

• Experience with audio language models, speech‑to‑speech, ASR, and TTS technologies.
• Familiarity with fine‑tuning, reinforcement learning, and preference alignment workflows.
• Knowledge of data‑generation for speech and dialogue systems.
• Understanding of API design and function‑calling mechanisms for voice‑driven applications.

What we offer

• Opportunity to shape applied audio AI at a foundation‑model company.
• Deep technical ownership and direct impact on enterprise deployments.
• Collaboration with a fast‑growing team at the intersection of research and product.