Staff ML Engineer (AI)

TL;DR

Staff ML Engineer (NLP/AI): Designing, training, and deploying fine-tuned NLP models for core product experiences including summarization and search ranking with an accent on the full model lifecycle from experiment to production. Focus on building scalable finetuning pipelines on GPU infrastructure and optimizing model architectures to serve millions of daily active users.

Location: Washington (Seattle), Georgia (Atlanta), or California (San Francisco)

Company

hirify.global is a global leader in CRM and cloud software, with this role specifically based in the Slack ML team.

What you will do

• Design and execute finetuning strategies for LLMs and other deep learning architectures tailored to summarization, ranking, and generation.
• Own the entire model training lifecycle, including data curation, infrastructure, hyperparameter optimization, and deployment.
• Build and maintain scalable finetuning training pipelines on GPU infrastructure.
• Collaborate with Product Managers, Designers, and Frontend Engineers to conceptualize and build new AI features.
• Lead high-impact multi-functional projects and contribute to technical architecture decisions.
• Mentor other engineers and improve engineering standards, tooling, and code review processes.

Requirements

• 5+ years of hands-on experience training and fine-tuning deep learning models in NLP or related domains (Speech, IR).
• 5+ years of experience with deep learning frameworks such as PyTorch, TensorFlow, or JAX.
• Proven track record of shipping fine-tuned models to production at scale.
• Proficiency in programming languages such as Python, PHP, Ruby, Go, C, Scala, or Java.
• Experience leading technical architecture discussions and driving strategic technical decisions.
• Strong communication skills and ability to write maintainable, testable code.

Nice to have

• Expertise with recommendation systems or search.
• Familiarity with model optimization for inference (quantization, pruning, TensorRT, ONNX).
• Experience with retrieval-augmented generation (RAG) and hybrid retrieval/generation systems.
• Knowledge of using structured, unstructured data, and knowledge graphs in RAG solutions.