Lead AI Engineer: Build Production-Grade Multi-Model Systems

Innovaccer Analytics; San Francisco (CA); On-site; USD 250,000 - 420,000; Full time

Benefits offered by this job

PTO benefits

Parental leave

Rewards & recognition

Insurance benefits

Job summary

Innovaccer in San Francisco seeks a Principal AI Engineer to lead production-grade AI systems at scale, taking ideas from paper to prototype to production. You will build the model layer of real products, selecting model sizes, composing multiple models, and meeting product-level accuracy goals.

You will design experiments, run distributed training with PyTorch, and ensure training runs are efficient. You will explain complex results to clinicians and operators, and contribute to research or

Qualifications

• MS or PhD in Computer Science, Machine Learning, or related quantitative field; exceptional BS candidates considered.
• First-author publications at NeurIPS, ICML, ICLR, ACL, EMNLP, or similar; meaningful open-source ML contributions; production-ready models.

Responsibilities

• Take an idea from paper to prototype to production.
• Build the model layer of a real product, choosing model sizes and composing models into a working system.
• Write real code in Python; understand training speed and systems implications.
• Design and run experiments with clear hypotheses and ablations.
• Measure results and build evaluation before modeling.
• Explain work to clinicians and operators who will use the outputs.

Skills

Python

PyTorch

Multi-GPU training

Distributed training

Model deployment

Experiment design

Communication with non-technical

Education

MS/PhD in CS/ML or related field

Tools

HuggingFace

DeepSpeed

vLLM

FSDP

Job description

Innovaccer in San Francisco seeks a Principal AI Engineer to lead production-grade AI systems at scale, taking ideas from paper to prototype to production. You will build the model layer of real products, selecting model sizes, composing multiple models, and meeting product-level accuracy goals.

You will design experiments, run distributed training with PyTorch, and ensure training runs are efficient. You will explain complex results to clinicians and operators, and contribute to research or