Staff Machine Learning Engineer, Homefeed

Description

The Retrieval team builds common ML tools that match people with ideas they will find useful.

This Staff Machine Learning Engineer role pushes that work forward using agentic AI.

Your main focus is Reflex, an effort to automate and speed up work on recommendation models.

You will own a major technical area and build agents that look into opportunities, make code changes, run experiments and learn from the outcomes.

You will work with other technical leaders and still write code on the hardest engineering and modeling problems.

You will also help with generative retrieval, mainly Semantic IDs: structured content representations that recommendation models can generate and use to fetch relevant items.

The role mixes applied research, production engineering and technical leadership, and it can improve both the recommendation models and how engineers build them.

Responsibilities

• Own a major technical area in automated recommendation research within Reflex, working with other tech leads and partner teams.
• Build LLM agents that generate hypotheses, implement them, run experiments, evaluate results and learn from earlier outcomes.
• Create rigorous evaluations of agent behavior and research output, and use them to improve quality, reliability and efficiency.
• Contribute to generative retrieval models, especially Semantic ID learning, evaluation and rollout into production recommendation systems.
• Move promising ideas from research and prototypes through offline evaluation, online experiments and production launches.
• Build dependable integrations with ML development infrastructure, balancing model quality, iteration speed, compute cost and operational complexity.
• Lift technical quality through hands-on coding, design and code reviews, documentation and mentoring.
• Work with research, recommendation and ML infrastructure teams to shape technical approaches, clear up dependencies and drive adoption.
• Find and apply useful advances from research, and build new methods where current approaches fall short.

Requirements

• Bachelor's degree in computer science, machine learning or a related field, or equivalent hands-on experience.
• At least 8 years building and shipping ML or AI systems, with substantial hands-on engineering and technical leadership.
• Proven experience building LLM agents or other agentic AI systems for complex tasks, and a clear method for measuring and improving them.
• Deep knowledge of machine learning basics, plus real depth in at least one area such as applied LLMs, representation learning, recommender systems, retrieval or ML systems.
• Background in turning research ideas into reliable production systems and measuring their impact with well-designed experiments.
• Strong software engineering skills: designing maintainable systems, building hard components and running thorough technical reviews.
• Experience leading a large technical area, deciding under ambiguity and working well across teams.
• A record of mentoring engineers and raising a team's technical standards.
• Clear written and spoken communication, including explaining technical tradeoffs to varied audiences.
• Careful, disciplined use of AI in your own work, including checking generated code, analysis and conclusions.

Conditions

US-based applicants only.

This is an individual contributor position; people management experience is not required.

The role can be based anywhere in the country, with in-person collaboration in the office 1-2 times per quarter.

The base salary range is $189,308 - $389,753 USD, and final pay depends on location, travel, relevant prior experience and specific skills.

Relocation assistance is not offered; details on the working model are on the company's flexible work page.

The company will teach you its stack and products and supports your growth as a technical leader.

The company aims for a fair, inclusive workplace for every employee.

Benefits

• Equity is available for this role.
• Details on culture and benefits are available on the company's website.