Designed agentic development environments with guardrails: tool calling, prompt/versioning, structured outputs, eval harnesses, and reliability checks (so “vibe coding” still ships safely).

Implemented end-to-end foundations for AI products: RAG/knowledge workflows, model routing, observability/telemetry, cost controls, and deployment—then iterated based on user feedback and real usage.

Note: Work under NDA; public-facing work and demos at olicheng.com.

Define product structure: flows, information architecture, interaction patterns, and requirements that engineers can build confidently.

Work hands-on with teams to implement web/mobile prototypes and ensure quality via component specs, edge cases, and testable acceptance criteria.

Use lean startup principles to uncover what matters, then drive prioritization, iteration, and product narrative.

Clinician-led discovery + rapid prototyping (UX team of one): Ran direct interviews with medical providers and their clients, built/iterated prototypes fast, and drove beta testing loops to validate workflows and usability before/after launch.

Multi-platform shipping + GTM bridge: Drove delivery across iOS, Android, and web while working under the CTO; partnered with business/marketing/engineering on pricing & packaging to align GTM with what the product could reliably deliver.

Built wireframes and prototypes; synthesized usability findings into concrete product changes.

Partnered with engineering to ship iterations under real constraints (edge cases, accessibility).

Bridged design and engineering teams to ship quickly and iterate from user feedback.

Applied usability testing/user interviews to improve real product flows.