AI Modeling Engineer: Voice & Multimodal (Stock Options)

Palona AI; Los Altos (CA); On-site; USD 140,000 - 210,000; Full time

Benefits offered by this job

Stock options

Benefits: medical/dental/vision/ret/le

Family leave

Disability protection

Paid time off & holidays

Learning and development

Job summary

Palona AI is seeking an applied AI Modeling Engineer to enhance the intelligence, safety, latency, and cost of Palona’s voice and multimodal agents in production environments. You will own model selection, routing, prompting, fine-tuning when justified, and evaluation methodologies, collaborating with product, engineering, and customer teams to move from hypothesis to reliable deployment.

You will build representative datasets, evaluate open models, and translate research into concrete product

Qualifications

• 3+ years of industrial experience in AI/ML or related technical domain.
• Strong Python and ML tooling foundations.
• Experience with production AI systems and model evaluation.

Responsibilities

• Develop modeling and experimentation strategies for voice, language, and multimodal agents.
• Build offline and online evaluations measuring accuracy, latency, cost, and user experience.
• Create datasets from simulations, annotation, and feedback while protecting sensitive data.
• Evaluate frontier/open-source models and decide build/buy/prompt/fine-tune decisions.
• Improve prompting, context construction, memory, tool-use policies, and model routing.
• Design and implement post-training methods when advantageous.
• Collaborate with speech/real-time engineers to reduce latency and improve ASR/TTS.

Skills

Industry experience

Python

ML foundations

LLMs

Experiment design

Evaluation pipelines

Tools

PyTorch

JAX

Hugging Face

Job description

Palona AI is seeking an applied AI Modeling Engineer to enhance the intelligence, safety, latency, and cost of Palona’s voice and multimodal agents in production environments. You will own model selection, routing, prompting, fine-tuning when justified, and evaluation methodologies, collaborating with product, engineering, and customer teams to move from hypothesis to reliable deployment.

You will build representative datasets, evaluate open models, and translate research into concrete product