← All speakers

Bio, Work & Ideas

Chris Hernandez

Conference affiliation: Chime · 2025

Chris Hernandez is senior manager of AI operations at Chime, where he develops the human oversight and operational safeguards behind automated customer service. His work positions frontline support employees as essential contributors to AI quality, evaluation, and accountability.

Hernandez began as a call-center agent before moving into quality management, machine learning, and natural-language processing. He worked with machine-learning teams at Apple, then joined Chime in a senior NLP quality role. He subsequently led speech analytics and advanced into AI operations within Chime’s member-experience organization.

His initiatives include automated contact-reason classification, AI-generated notes, and Jade, Chime’s voice AI. He favors automating repetitive interactions while preserving straightforward access to human support for sensitive or complicated situations.

  • Human review as continuous feedback. Hernandez uses human corrections to identify hallucinations, refine evaluation criteria, and improve automated responses, prioritizing oversight where decisions carry greater risk.
  • Frontline teams as AI quality operators. Customer-experience and quality-assurance specialists contribute domain expertise through prompt testing, output labeling, edge-case identification, evaluation datasets, and ongoing performance monitoring. His AI Engineer conference session makes the case for involving these teams throughout development.
  • Operational playbooks for customer-facing AI. Hernandez emphasizes prohibited-response checks, clearly assigned incident ownership, escalation deadlines, tracked remediation, and recurring review. His approach to designing for operational failures treats customer impact and employee workload as central measures of whether automation succeeds.

Read the topics behind these talks

1 conference talk

Key ideas

Scroll to read ↓

A promising AI product needs more than a launch: evaluations, human feedback, and clear quality ownership turn rapid development into reliable improvement.

  • What happens after the first version ships?
    0:00 ↗
  • Faster development increases operational demands
    1:12 ↗
  • Crossing the quality chasm
    2:09 ↗
  • Confident errors need human judgment
    3:34 ↗
  • Feedback needs reviewers
    5:09 ↗
  • Expanding QA beyond retrospective audits
    6:52 ↗
  • Give AI quality an owner
    9:09 ↗
  • Place review where it matters, then keep operating
    11:13 ↗

References