17 Nov
|
JoinStartup
|
Winnipeg
17 Nov
JoinStartup
Winnipeg
Define conversational AI quality remotely from Canada or the United States using Python, LLM APIs, and rigorous evaluation systems.
As Cresta’s mid-level AI Agent Evaluation Engineer, you will own quality measurement for production LLM-based conversational agents. You will combine conversation analysis, annotation, rubric design, and statistical sampling to create an accurate picture of agent performance. Your work will support Cresta’s AI Receptionist, which handles millions of conversations monthly across hundreds of customers.
Key Responsibilities
Define rubrics across adherence, outcome, and conversation experience
Review production conversations and classify failure modes weekly
Build automated judges, golden sets, and regression suites
Calibrate judges against human labels and measure agreement
Report prioritized quality findings to engineering and product owners
Requirements
Three or more years in software engineering, applied ML, or research
One to two years working with production LLM systems
Python programming skills and experience with LLM APIs
Experience building end-to-end evaluation systems
Degree in a related field or equivalent practical experience
Bring Python, annotation, rubric design, and LLM evaluation expertise to measure Cresta agent quality accurately.
📌 AI Agent Evaluation Engineer (Winnipeg)
🏢 JoinStartup
📍 Winnipeg