Data Analytics Lead

present of AI; San Francisco (CA); Sur place; USD 120 000 - 170 000; Plein temps

Résumé du poste

HappyRobot is seeking an analyst to turn fuzzy product questions into clear measurements and tests for AI agents that handle calls and emails. You will design experiments, apply statistics, and use SQL and Python to quantify performance of voice and email agents in real operations.

You will link lab scores to customer outcomes, analyze regressions on large conversation datasets, choose metrics that reflect real value, and communicate findings to engineers and non-technical stakeholders.

Qualifications

• Solid statistics background.
• Experience with experiment design and causal thinking.
• Strong SQL and Python skills.

Responsabilités

• Measure AI agents that handle calls and emails with rigorous metrics.
• Link lab model scores to real customer operations and analyze regressions across large conversation data.
• Decide which metrics reflect real value and persuade the team to act on them.
• Explain findings to engineers and non-technical stakeholders.

Connaissances

Statistics

Experiment design

Causal thinking

Explain findings

Outils

SQL

Python

Description du poste

The role asks for someone who can turn fuzzy product questions into clear measurements and tests for AI agents that handle calls and emails. A strong candidate brings solid statistics, experiment design and causal thinking, plus strong SQL and Python. They also have the judgment to tell a statistically real result from one that matters to a customer, and can explain findings to both engineers and non-technical people.

How it fits the AI era

AI agents give different answers each time, so measuring them takes more care than testing a normal software feature. The analyst links lab model scores to what happens in real customer operations, and digs into regressions across huge volumes of conversations. Deciding which metrics reflect real value, and persuading a team to act on them, stays a human job.

How it builds AI experience

Working here would mean measuring how changes to models, prompts and workflows affect voice and email agents running real operations for enterprise clients. To build this experience beforehand, someone could practice evaluating an LLM app, comparing offline test scores with live results, and reviewing transcripts by hand alongside the numbers.

Worth showing

• A/B testing and causal inference
• Advanced SQL and Python
• Defining product and agent metrics
• Root-cause analysis of regressions
• Explaining findings to non-analysts

This is our own read, written from the employer's posting and the company's Spotlight. The employer's posting is the source of truth for the role.

About HappyRobot

AI voice and email agents that autonomously handle the phone calls and operational busywork of supply chain and enterprise teams.