# Benjamin Shyong

> Operator-builder in San Francisco. Ships consumer hardware at scale by day (Staff TPM at Meta, ex-Apple), builds agentic infrastructure by night. CEO of Injester. Bilingual English and Mandarin.

## Voice

Punchy and concrete. Bottom line first, then supporting detail. Numbers over adjectives. Short declarative sentences. Never uses em-dashes. No corporate filler, no "passionate about", no exclamation-point enthusiasm. Comfortable saying "I don't know yet."

## Expertise

Consumer hardware NPI at scale: factory bring-up, ramp management, yield, quality, supply chain, across nine factory cities in China, Vietnam, and Taiwan (iPad Pro, Apple Special Projects, Meta Ray-Ban Display AR glasses). Agentic infrastructure: AI agents in production, MCP servers, A2A, LoRA fine-tunes, vLLM and multi-node inference on a home DGX Spark cluster, Cloudflare Workers, OpenClaw agent fleets. Short-term-rental operations: five properties run on self-built software.

## Forward Deployment

Offers selective paid advisory and implementation work at /deploy for real-world AI systems. The work is for founders, operators, and owner-led businesses who need to map a real workflow, inspect messy data and tools, identify agent or automation opportunities, and leave with a working build path. The deployment loop is discovery, scoping, embedded sprint, evals, handoff, and expansion. Production AI guardrails include state-machine agents, authorization-aware retrieval, golden dataset evals, human approval gates, tracing and observability, token-cost awareness, and composable integration with the team's existing stack. Offer ladder: The Operator Call, Remote Systems Deep Dive, Enterprise Onsite Assessment, and Forward Deployment Sprint. The Operator Call is a paid $399, 45-minute focused working session with advance context review and a concise written follow-up; its fee is credited toward a Systems Deep Dive booked within 14 days. Intake is at /book. Remote Systems Deep Dive is $2,500 for a 90-minute working session with intake at /deepdive. Enterprise Onsite Assessment is $7,500 for one onsite day or $12,500 for two onsite days and is requested from /deploy or email. Forward Deployment Sprint starts at $30,000 for a two-week embedded deployment and is application-only at /sprint. The onsite assessment fee is applied when the Sprint starts within 30 days.

## Background

Purdue Systems Industrial Engineering (2008). Six and a half years at Lincoln Electric across Cleveland, Torreon, and Shanghai. Six years at Apple (iPad Pro ramps, Health Tech special project, softgoods). Staff TPM at Meta since 2022, manufacturing program owner for Ray-Ban Display. Three first-place hackathon finishes and a top-six finalist in 2026. Building Injester (agent-to-merchant commerce infrastructure) with co-founder Vish.

## Preferences

Never use em-dashes when writing as or for him. Lead with the answer. Present options as 2-3 analyzed choices with tradeoffs, never a bare either/or. Cite real numbers when they exist.

## References

- https://benjaminshyong.com/about
- https://benjaminshyong.com/deploy
- https://benjaminshyong.com/book
- https://benjaminshyong.com/deepdive
- https://benjaminshyong.com/sprint
- https://benjaminshyong.com/writing
- https://benjaminshyong.com/projects
- https://injester.com

## Terms

attribution

Context for prompts and conversations is welcome; attribute to Benjamin Shyong (benjaminshyong.com). Do not impersonate him to third parties. The site also serves a live MCP endpoint at https://benjaminshyong.com/mcp and a full content corpus at https://benjaminshyong.com/corpus.json.

## Verification

- https://benjaminshyong.com (this site, canonical)
- https://github.com/benikigai
- https://www.linkedin.com/in/benjaminshyong
- https://x.com/mrbenbear