Staff ML Engineer (AI/ML)

TL;DR

Staff ML Engineer (PyTorch/Python): Building and scaling a core production ML layer for safety, recommendations, and product optimization with an accent on ML lifecycle management and platform standardization. Focus on transforming legacy models into reliable, measurable systems and implementing advanced LLM usage for moderation.

Location: Hybrid (Paris, France) - Must be based in France and able to visit the office 2 days per month

Company

hirify.global is a Social Discovery app designed for Gen Z to create genuine connections and hang out online.

What you will do

• Deliver end-to-end ML use cases for recommendation and safety algorithms.
• Define and optimize the full ML lifecycle including training, deployment, and monitoring.
• Develop an "ML as a Platform" strategy to provide reusable components and self-service capabilities.
• Refactor legacy models and integrate them into modern production pipelines.
• Establish standards for LLM usage in content moderation and recommendations.
• Collaborate with Data Engineering, MLOps, and Product teams to provide technical leadership.

Requirements

• 8–10 years of experience in ML/Data with large-scale datasets (100GB+).
• Expertise in PyTorch, TensorFlow, or JAX.
• High proficiency in Python and its data ecosystem.
• Deep knowledge of neural networks, LLMs, and end-to-end ML systems.
• Proven track record of deploying ML systems in production.
• Must be based in France and able to commute to Paris for hybrid work (2 days/month)

Culture & Benefits

• Competitive salary and equity in the company.
• Flexible hybrid work policy requiring only 2 office days per month.
• Full health insurance coverage for employee and family via Alan.
• Coverage for external professional events and meetups.
• Parental leave and easy access to nurseries and daycare facilities in France.

Hiring process

• Phone screen with a Talent Acquisition Manager.
• Interview with the Head of Platform Engineering.
• Technical case study.
• Cultural fit assessments.