Angelo Zhang
Full-stack & platform engineer · Multiple AI products, one stack
Beijing · Remote-friendly
Twelve years of consumer software. Currently building and operating production AI products as an independent engineer.
zhangmodel825@gmail.com·www.linkedin.com/in/angelo-zhang-0987a7409
Open to staff IC, founding engineer, and architect conversations
I am a hands-on full-stack engineer and independent product builder. I take AI products from blank file to live operation — product judgment, system architecture, full-stack execution, release, and ongoing care of the thing after launch.
My strongest work is turning ambiguous early ideas into systems that survive contact with real users: deciding what is worth building, shaping the boundaries, shipping the first version, keeping it alive long enough to learn what is actually worth doing next.
I increasingly build platforms that host multiple products rather than building one product at a time. The same monorepo, shared libraries, shared service interfaces, and shared knowledge base now power several public AI products with a team of one — turning the marginal cost of trying a new product from months into weeks.
Before going independent, I spent five years building consumer Android applications at Sogou, Qunar, and Ctrip. That mobile-era discipline — release quality, performance, the user with a slow phone in a quiet city — still shapes how I work.
Now
Several production AI products are in active operation. I split my time between product judgment for the next quarter, the unglamorous infrastructure work that lets a one-person team feel like a small team, and a quiet rewrite of the backbone they share.
Open to staff-level or founding-engineer conversations where the bar is shipping, not slideware.
Core strengths
0-to-1 product delivery
Turns vague product direction into shipped, usable systems with a clear path for iteration.
System-level architecture
Balances speed with long-term maintainability so systems can be launched, operated, and evolved.
Production ownership
Owns release readiness, recovery paths, observability, and ongoing quality — not just the first commit.
Consumer product craft
Carries mobile-era instincts for reliability, interaction quality, performance, and disciplined releases.
Capabilities
AI product architecture
Turning AI capabilities into product systems that hold up under real users, not demo videos.
Multi-layered memory · multi-provider LLM orchestration · stream-first chat UX · tenant-isolated billing rules
0-to-1 delivery
Closing the long tail from working prototype to released product — release governance, recovery, drift, the work most plans under-budget.
Local CI parity · declarative expand/contract migrations · idempotent retries · observable failure paths
Full-stack execution
Moving fluidly from interaction craft to backend services, data shape, and live operating paths.
TypeScript end-to-end · React Server Components · long-running async jobs · webhook + cron recovery
Engineering discipline
Treating boundaries, isolation, and replay-safety as first-class product properties, not afterthoughts.
Auth boundaries at the edge · tenant-scoped data · feature-flagged rollouts · environment drift detection
AI-native operating model
Treating AI as a first-class collaborator: MCP interfaces replace internal dashboards, a code-synced knowledge base feeds product and engineering AI work, and shared infrastructure makes additional products cheap to launch.
MCP tool surfaces · code-synced KB repository · multi-product monorepo · shared platform libraries
Selected work
LivelyVerse— A public AI product taken from product direction to launch and live operation as an independent builder.
Founder · 2024 — present · livelyverse.com
Approach — Multi-layered memory · multi-provider LLM orchestration · stream-first chat · tenant-isolated billing rules
- Own product positioning, user experience, system design, engineering delivery, launch readiness, and live operations.
- Compress complex capabilities into a product surface that is understandable to users and still has room to evolve.
LeyLine Pro— A second public AI product built with the same lean, independent 0-to-1 operating model.
Founder · 2024 — present · www.leylinepro.ai
Approach — Long-running media generation · provider failover · idempotent jobs · resilient credit accounting
- Moved from concept, pages, and core flows to public release as a solo engineer-founder.
- Use real feedback to refine positioning, interaction details, product boundaries, and delivery quality.
Mendr— AI-native developer tooling that diagnoses, fixes, tests, and verifies engineering issues end-to-end as one loop.
Founder · 2025 — present · mendr.dev
Approach — AI agent orchestration · multi-step task execution · structured tool calls · test-and-verify feedback loops
- Bring AI-driven issue resolution to the developer-facing surface of an engineering team in a single loop.
- Apply the 0-to-1 discipline of consumer AI work to a developer-tools product surface.
Migo Read— Leveled English reading for young learners — AI-generated, age-appropriate content shaped into a graded curriculum, not raw model output.
Founder · 2025 — present · www.migoread.com
Approach — AI content generation pipeline · level-graded curation · interactive reading UX · learner progress modeling
- Apply AI content generation to consumer language education for young learners.
- Translate model output into a shaped curriculum suitable for graded learners, with care for the audience.
Experience
- 2024 — Present
Independent Engineer · Founder — Self-employed
Building and operating public AI products independently, with full ownership from product definition and system design to live service.
- Shipped LivelyVerse, LeyLine Pro, Mendr, and Migo Read from blank slate to public websites and usable product surfaces.
- Own product, design, engineering, launch, support, and iteration without a large team, turning early ideas into operated products.
- Prioritize durable quality, clear boundaries, disciplined releases, and real feedback over one-off technical spectacle.
- 2018 — 2023
Independent Product Engineering — Various projects
Expanded from mobile specialist into full-stack product engineering across product, interface, service, deployment, and operation.
- Built the breadth needed to connect product judgment, interface implementation, service workflows, deployment, and operating feedback.
- Developed a pragmatic builder style: small releases, direct feedback, clear tradeoffs, and ownership of the final outcome.
- 2015 — 2018
Senior Android Engineer — Qunar · Ctrip
Worked on mature consumer travel applications where reliability, performance, and release quality mattered.
- Delivered Android product work in cross-functional environments with strong QA and release expectations.
- Built long-term instincts for production stability, performance, and polished mobile interaction details.
- 2013 — 2015
Android Engineer — Sogou
Started professional engineering career building consumer Android application features.
- Learned to ship within established product and engineering standards early in a large-company environment.
Education
- 2009 — 2013Southwest University of Science and Technology — Bachelor's degree, Communication Engineering