AI Engineer, Agents – Evaluation
Guild.ai
full-time
Posted on:
Location Type: Hybrid
Location: San Francisco • California • United States
Visit company websiteExplore more
Tech Stack
About the role
- Create Task Evaluations That Matter: Design and implement task-specific evaluations that measure and improve agent quality. Each eval should both drive concrete iteration on our agents and spark broader innovation around the task itself.
- Define Tasks, Datasets, and Harnesses: Clearly specify tasks, collect and curate balanced datasets, and build robust evaluation harnesses that can be used across agents and modeling approaches. There is ample room for architectural design and systems thinking here.
- Build and Use a Reusable Evaluation Framework: Develop frameworks and tools for running evaluations at scale. Use these frameworks to tune existing agents and to guide the development of new ones in our environment.
- Explore Agent Orchestration Strategies: Investigate and implement orchestration patterns (tooling, routing, decomposition, multi-agent setups, etc.) that allow agents to tackle increasingly complex, multi-step, and long-horizon tasks.
- Apply Post-Training Techniques: Experiment with post-training approaches (e.g., fine-tuning, preference optimization, reward shaping, distillation) to produce high-performance models tailored to specific tasks and workflows.
- Run Experiments End-to-End: Design, run, and analyze experiments with rigor. Turn experimental results into clear recommendations and concrete changes to model configurations, prompts, and system design.
- Collaborate Deeply Across the Stack: Work closely with founders, product, and infrastructure engineers to ensure evaluations, agents, and platform primitives all reinforce each other.
Requirements
- MS or Ph.D. in a relevant field (e.g., Computer Science, Machine Learning, NLP) or equivalent practical experience
- Strong background in machine learning and large language models, ideally including both research and hands-on implementation
- 2–5 years working with LLM technology, with familiarity across:
- Prompting and interaction patterns
- Agent and tool orchestration strategies
- Evaluation strategies for complex, open-ended tasks
- Proficiency writing production-quality code, especially in Python; comfort working with TypeScript or modern web/backend stacks
- Experience designing and running experiments, and interpreting results in messy, real-world settings
- Self-motivated, comfortable operating in an unstructured, high-ambiguity environment
- Strong communication skills and the ability to translate vague goals into concrete, testable setups.
Benefits
- Significant equity in an early-stage, venture-backed startup
- Comprehensive Health Benefits (Medical, Dental, Vision)
- Flexible PTO to ensure you have the time you need to recharge
Applicant Tracking System Keywords
Tip: use these terms in your resume and cover letter to boost ATS matches.
Hard Skills & Tools
machine learninglarge language modelsPythonTypeScriptevaluation strategiespost-training techniquesexperiment designdata curationarchitectural designsystems thinking
Soft Skills
self-motivatedstrong communicationcollaborationproblem-solvingadaptabilityanalytical thinkinginnovationrigortranslating goalsoperating in ambiguity
Certifications
MS in Computer SciencePh.D. in relevant field