Drimva

Scale AI

Senior/Staff Machine Learning Research Engineer, General Agents, Enterprise GenAI

$290K–$363K

San Francisco, CA +4 · Staff · Full-time · Posted Feb 10, 2026

Location
San Francisco, CA
New York, NY
New York, New York, United States
San Francisco, California, United States
Seattle, Washington, United States
Experience
5+ years building and deploying machine learning or AI systems for real-world production use cases.
Education
Bachelor’s and/or Master’s degree in Computer Science, Machine Learning, AI, or equivalent practical experience.

About the job

Designs, builds, evaluates, deploys, and improves production-ready AI agents for recurring enterprise use cases.

What you'll do

  • Build end-to-end agent systems and scalable architectures integrating LLM reasoning, tools, memory, and control logic.
  • Create evaluation frameworks and measure agent performance, reliability, and business impact; monitor deployments and analyze failures.
  • Collaborate with customers and cross-functional teams to translate enterprise needs into robust agent systems and contribute to technical direction.

What they ask for

  • 5+ years building and deploying machine learning or AI systems for production use cases.
  • Bachelor’s and/or Master’s degree in Computer Science, Machine Learning, or AI, or equivalent practical experience.
  • Deep understanding of modern LLMs, prompt-, context-, and system-level optimization, and agentic system design.
  • Proficiency in Python and experience integrating models with external tools, APIs, databases, and services.

Nice to have

  • Hands-on experience building AI agents with generative AI stacks; experience with agent frameworks, orchestration, or workflow systems.
  • Experience with production LLM evaluation and observability, cloud deployment at scale, or adapting foundation models using SFT, RLVR, or LoRA.

Pay

$290,400 — $363,000 USD (base salary)

Tools

Python, LLMs, OpenAI APIs, SFT, RLVR, LoRA

AI focus

Develops production enterprise agents using LLM reasoning, tool use, memory, planning, multi-step reasoning, and multi-agent patterns; evaluates and improves their reliability and real-world impact.

Model-written summary · the original posting is authoritative

First seen Oct 10, 2026 · Last verified Nov 5, 2026 (UTC)