Staff ML Engineer — Memory & Agent Retrieval

Zep AI; San Francisco (CA); On-site; USD 180,000 - 260,000; Full time

Job summary

Zep AI in San Francisco seeks an engineer who builds agents and trains the models they run on. You will own the memory extraction and retrieval loop, finetune models, and push systems to production. You’ll report to the founder and join a team with experience at Scale AI, Dropbox, and McKinsey.

You will work in a distributed team, write production-grade Python, design experiments, and build eval harnesses to catch regressions before release. A Master’s in CS or equivalent is required.

Qualifications

• Shipped production ranking, retrieval, or query understanding.
• Built production ML solutions for memory-extraction problems.
• Strong Python and ML tooling experience.

Responsibilities

• Build agents that improve retrieval, including query understanding and ranking.
• Fine-tune models for memory extraction and production deployment.
• Develop and maintain the evaluation harnesses to detect regressions.
• Own end-to-end experiments from dataset creation to production launch.
• Document findings and share results with the team.

Skills

Production retrieval

Transformer architectures

Python

Experiment design

Education

Master's in Computer Science or equivalent experience

Tools

PyTorch

AWS

SQL

Job description

Zep AI in San Francisco seeks an engineer who builds agents and trains the models they run on. You will own the memory extraction and retrieval loop, finetune models, and push systems to production. You’ll report to the founder and join a team with experience at Scale AI, Dropbox, and McKinsey.

You will work in a distributed team, write production-grade Python, design experiments, and build eval harnesses to catch regressions before release. A Master’s in CS or equivalent is required.