Machine Learning Engineer (AI)

Location: Remote - Americas

Company

hirify.global is an e-commerce platform helping entrepreneurs and enterprises build and scale businesses for merchants in 175 countries.

What you will do

• Develop and deploy generative AI, natural language processing, and HSTU-based recommendation models at scale.
• Design scalable AI/ML system architectures and sophisticated inference pipelines processing billions of events for real-time recommendations.
• Build data pipelines for model training, fine-tuning, and evaluation across merchant events, consumer interactions, and payment sequences.
• Experiment with novel architectures and optimize production systems using negative sampling, ANN search, and distributed GPU training.
• Collaborate with product teams, data scientists, and infrastructure engineers to deliver measurable business impact.
• Translate complex machine learning concepts into actionable insights for technical and non-technical audiences.

Requirements

• Mastery of recommendation systems, generative AI, or LLMs.
• End-to-end experience training, evaluating, testing, and deploying machine learning products at scale.
• Experience building data pipelines and making ETL design decisions using disparate data sources.
• Proficiency in Python, shell scripting, streaming and batch data pipelines, vector databases, DBT, BigQuery, BigTable or equivalent technologies, and orchestration tools.
• Experience running machine learning in parallel environments, including distributed clusters and GPU optimization.
• Availability for on-call work may be required.

Culture & Benefits

• Digital-first daily work environment.
• Fast-paced work with significant change, complexity, and ambiguity.
• Emphasis on critical thinking, resilience, resourcefulness, and constructive disagreement.
• Focus on using AI agents and tools for suitable tasks while prioritizing human judgment.

Hiring process

• Candidates should be prepared to interview within one week.
• The interview loop is intended to be completed within 30 days.
• The process includes a pair-programming interview using the candidate's own IDE.