Agent memory, built from scratch
Designed and built the file-based agent memory system and continue to improve it. Exploring memory as an entry point for improving the agent harness itself, so experience can inform how agents approach future work.
Building at CREAO AI
📍 Fremont, CA
Arky Yang
Tech Lead — Agent Core & Creao Platform at CREAO AI.
I build the agent and the platform around it, working across engineering, product management, and go-to-market to turn capabilities into useful products. I help teams align on what we’re building and why, and lead junior engineers in strategy, research, and delivery.
01 — Summary
Build the systems. Understand the people using them.
I lead development of Agent Core & Creao Platform, bringing deep technical understanding to product management. I connect engineering and go-to-market to shape product direction and help engineers build their product skills. Previously at Atlassian, I took an enterprise context engine from zero to external release across 10+ partner teams.
Platform
Hands-on development of the agent core and the platform around it
Product
Connecting market needs and technical capabilities to shape product direction
Partnerships
Working with technology partners to unlock customer workflows
Leadership
Mentoring engineers across strategy, research, and delivery
02 — Current Work
Tech Lead — Agent Core & Creao Platform
Connector architecture & strategy
Restructured the connector architecture and brought on exclusive connectors that enable agents to do real work for users. Established an agent-led process combining market analysis with engineering cost to prioritize connectors, their capabilities, and the capacity to deliver them.
Agent-driven connector maintenance
Built a maintenance workflow using CREAO’s own agents to monitor connector issues, prioritize failures, and create automated agent tasks to fix them. The goal is to reduce ongoing maintenance across a heterogeneous ecosystem of integrations and changing dependencies.
Technical partnerships that unlock customer workflows
Led the TinyFish integration to make web access a default agent capability. Collaborate with TinyFish’s engineers and co-founders to shape integrations around customer workflows and address gaps using their broader platform.
Read the integration story ↗Product management & team leadership
Take on product management responsibilities in an organization without dedicated PMs. Bring go-to-market insights and engineering knowledge together to define product direction, priorities, and scope for team features and the platform’s overall user experience. Lead and mentor junior engineers in strategy and engineering planning, exploring user and market research together in the AI field. Build a shared understanding across both teams and help engineers develop the skills to shape the product themselves.
03 — Previously
Software Engineer, Code Intelligence
Enterprise context engine, from zero to external release
Built the retrieval agent that plans search, exploration, and file reads to assemble context for downstream AI features. Designed the cross-repo code search API, exposed it through MCP and a CLI agent, and made querying and indexing pluggable across providers and ranking algorithms.
External onboarding across 10+ partner teams
Coordinated source control, provisioning, billing, gateway, and AI product teams to deliver an end-to-end onboarding journey for external customers. Took code search from one source host to multiple hosts, with tenant-aware permissions and documented service-to-service integration contracts.
Production systems & user-facing experiences
Delivered rate-limit policies, traffic attribution, and tracing alongside React/TypeScript code search and AI chat interfaces. Improved the team’s agent workflows with a Splunk MCP for debugging, Slack support automation, and a shared context registry.
04 — Side Projects
Personal projects
🎙
Shipped · macOS
AnyScribe
Local, on-device meeting transcription for your Mac.
A menu-bar app that live-transcribes meetings — capturing both your mic and the Mac's
system audio and labeling them Me / Them. Everything runs
on-device with whisper.cpp
(Metal-accelerated). No audio ever leaves your machine.
- One hotkey to start; live
Me/Themtranscript - Auto-saves timestamped Markdown to a folder you choose
- Chinese / English / mixed code-switching
- Signed & notarized — opens with no Gatekeeper warnings
Native Swift · whisper engine bundled — no extra setup.
🛠
In progress · unreleased
Rederive
A "Roblox for education" platform with an agent harness for creation.
Anyone with an AI agent can author courses in their domain, and a Socratic teaching agent (also built by me) guides learners through them. The authoring side is a BYO-agent pipeline — fully autonomous (YOLO) or human-in-the-loop — that goes from one prompt to a fully animated, narrated video or interactive course.
- Autonomous long-horizon execution
- BYO-agent contracts — a packaged plugin bundling skills + MCP
- Zero-to-one creative pipelines
- The pedagogy & UX of a teaching agent
Same problem shape as agent-coding harnesses — applied to education.
🐻
Shipped
Cedartale
Voice-cloned AI storytelling app for kids.
A parent's own voice can read the story. Founder, sole engineer, sole product designer — the whole loop from idea to a shipped consumer app, including user research. In 3 weeks.
- User research with parents & kids
- Self-deployed voice-cloning & caption-alignment models
- One-click picture-book generation with theme & style consistency
- Consumer product, end to end
Founder · sole engineer · sole product designer.
05 — Stack
Tools & technologies I work with
Backend
Kotlin / SpringPythonGo
RedisPostgresGraphQLOpenSearch
RailwayModal
Infra & platform
Edge / API gatewayService-to-service auth
User authEvent-streaming
Container deploy pipelines
AI
CodexClaude CodeRovodev CLI
CursorWindsurf
OpenClaw + GLMMCPs & skills
Observability
Distributed tracingProduction log platforms
SplunkSignalFxTime-series dashboards
Frontend
ReactTypeScript
Admin & portal UIs
06 — Education & before
Chemistry, then computer science
MS, Computer ScienceUT Austin
Chemistry, Resnick FellowCaltech
BS, Chemistry (Highest Honors)UC Berkeley
ChemistryHKUST
Visiting Student, Organic ChemistryUC Santa Barbara
CS coursework: Android, deep learning, computer vision, optimization, front-end. Pre-CS research: 2D organic–inorganic hybrid perovskite materials for optoelectronics (Harry Atwater Group, Caltech), copper nanowire synthesis for transparent electrodes (Peidong Yang Group, UCB), polymer solar cells (He Yan Group, HKUST). Publications include Giant Enhancement of Photoluminescence Emission in WS2–Two-Dimensional Perovskite Heterostructures.
Earlier — SWE Intern (2021)
- React/TypeScript + Jest reusable card UI components for distributed services.
- Event-driven Spring Boot/Kotlin/GraphQL backend for default-component provisioning on site creation.
- In-product help search experiments (Algolia, React Context & Hooks) with content design.
Outside engineering (it shows up in the work)
- President, Caltech Entrepreneurship Club (2017–2020) — alumni-CEO panels, the FLOW/Rockey Fund program for young tech companies, Idealab field trips.
- Plug and Play (2019) — analysed pitch presentations across verticals.
- Calibr at Scripps Research (2018) — research & market landscape of new depression meds.
- Mom of a two-year-old — an excellent forcing function for thinking carefully about AI.