Senior ML Platform Engineer - Remote

Neara; United States; On-site; USD 152,000 - 228,000; Full time

Benefits offered by this job

Flexible Time Off

Health Benefits

401(k) match

Stock Options

Mental Health Program

Paid Holidays

Job summary

Invoca is seeking a Senior ML Engineer to own the production ML stack, deploying high-throughput inference and reliable services for product teams. You will optimize pipelines, monitor performance, and drive cost efficiency.

You will collaborate with Data Scientists and Applied AI Engineers to ship AI-powered features, continuously improving model serving and deployment in production.

Qualifications

• 5+ years of ML Engineering with production focus.
• Advanced Python and deep learning proficiency (PyTorch, Transformers).
• Experience deploying transformer-based NLP models in production.
• Fine-tuning SLMs/LLMs (LoRA, QLoRA, PEFT) and optimization for throughput.
• Experience with inference infrastructure: Triton, Baseten, vLLM, TGI, SageMaker, Vertex AI.

Responsibilities

• Own production ML as a system with CI/CD for ML artifacts.
• Provide operational excellence for ML stack: uptime, latency, throughput, cost.
• Own inference infrastructure end to end: model serving on Triton/Baseten/Kubernetes.
• Fine-tune models and connect decisions to customer outcomes.
• Collaborate with data scientists and product teams to ship features.

Skills

Python proficiency

PyTorch

Transformers

Fine-tuning SLMs/LLMs

Inference optimization

API development

MLOps tooling

RLHF knowledge

CS/Engineering background

Agentic AI workflows

Education

B.S. in Computer Science or related

Tools

Triton

Baseten

vLLM

TGI

SageMaker

Vertex AI

Job description

Invoca is seeking a Senior ML Engineer to own the production ML stack, deploying high-throughput inference and reliable services for product teams. You will optimize pipelines, monitor performance, and drive cost efficiency.

You will collaborate with Data Scientists and Applied AI Engineers to ship AI-powered features, continuously improving model serving and deployment in production.