Embedded ML Engineer: End-to-End On-Device AI Pipelines

Doist; San Francisco (CA); On-site; USD 140,000 - 220,000; Full time

Benefits offered by this job

Health insurance

401(k) matching

Unlimited PTO

Equity

Job summary

Liquid AI is seeking a hands-on ML Engineer to own the end-to-end model development pipeline for our on-device, automotive-focused projects. You will translate partner specs into concrete training requirements and manage data generation, model fine-tuning, and evaluation across software releases.

You will work directly with embedded teams and partner product managers to deliver production-ready checkpoints, with a focus on language coverage, tool-use fidelity, and robust dashboards for

Qualifications

• Hands-on ML experience: ~2+ years, strong internship track accepted.
• Trained models end-to-end in any modality (vision, audio, LLMs).
• Experience with large-scale data pipelines and data wrangling.
• Automotive/on-device ML context, first-principles reasoning.
• Strong communication skills; client-facing role.

Responsibilities

• Translate broad feature specs into concrete model training requirements.
• Own the core fine-tuning recipe for an on-device model; ensure tool calling accuracy across languages.
• Generate, clean, and analyze training data; build and maintain large-scale data pipelines.
• Run training and evaluation cycles across major software releases.
• Create dashboards, analyses, and polished partner-facing deliverables.
• Progressively own end-to-end model development from spec intake through delivered checkpoint.

Skills

Hands-on ML

End-to-end model training

Large-scale data pipelines

Automotive/on-device ML

Strong communication

Job description

Liquid AI is seeking a hands-on ML Engineer to own the end-to-end model development pipeline for our on-device, automotive-focused projects. You will translate partner specs into concrete training requirements and manage data generation, model fine-tuning, and evaluation across software releases.

You will work directly with embedded teams and partner product managers to deliver production-ready checkpoints, with a focus on language coverage, tool-use fidelity, and robust dashboards for