← All speakers

Bio, Work & Ideas

Dan Feng

Conference affiliation: Senior Director of Engineering · Maven Clinic · 2026

Dan Feng is an engineering leader specializing in generative AI for digital healthcare. As Senior Director of Engineering at Maven Clinic in 2026, he led the company’s AI platform and core member experience, applying AI to healthcare services where errors can affect appointments, benefits, and reimbursements.

His work included Maven Intelligence, an AI-powered healthcare orchestration layer connecting clinical context, member history, benefits information, and professional oversight across Maven’s virtual clinic. He also helped expand internal adoption of tools including Cursor and Claude Code, emphasizing shared infrastructure that supports employees beyond early adopters.

Distinctive operating principles

  • Product-minded engineering ownership: Feng expects engineers to understand member needs, solve ambiguous problems independently, and exercise architectural judgment as AI automates implementation.
  • Short planning cycles: He favors concise planning documents and two-to-four-week delivery cycles, arguing that changing model capabilities make detailed quarterly roadmaps unreliable.
  • Accountable code review: To manage increased AI-generated output, he supports smaller, stacked pull requests and self-merging straightforward changes with clear individual responsibility. Superficial approvals, he argues, create dangerous false confidence.
  • Risk-tiered AI reliability: Recoverable appointment-booking failures require different controls from reimbursement errors. For consequential workflows, he uses multi-model verification, checks whether systems agree, and escalates unresolved cases to human support.
  • Continuous evaluation with human oversight: Repeated integration tests, rubric-based conversation assessments, and manual review help teams detect inconsistent model behavior and refine safeguards after release.

Feng’s AI Engineer World’s Fair presentation also outlines a longer-term goal: AI-assisted design, implementation, deployment, production monitoring, and remediation, with human accountability maintained wherever healthcare or financial risks demand it.

Read the topics behind these talks

1 conference talk

References