Skip to content
View roystonlmq's full-sized avatar
0 followers · 3 following

Block or report roystonlmq

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
roystonlmq/README.md

Royston Lee

Customer solutions consultant building reliable agentic AI systems

I turn ambiguous customer workflows into systems that can be deployed, evaluated, and trusted. My work sits between customer discovery, solution architecture, implementation, and adoption—especially where AI agents need clear permissions, evidence, recovery paths, and measurable outcomes.

Based in Singapore and currently working on agentic customer-experience systems at Ada. Previously led more than 200 enterprise AI deployments across APAC.

LinkedIn · Singapore

Selected work

Project Product and engineering evidence Stack
Career Evidence Engine Grounds tailored career claims in source evidence, exposes unsupported requirements, and enforces human review. Python, pytest, DOCX
Agentic Travel Planner Demonstrates grounded proposals, confirmation-bound writes, idempotency, exact receipts, and scoped undo. React, TypeScript, Vitest
Customer Evidence Inbox Captures screenshots and AI-assisted notes in a multi-user customer-context workflow. React, PostgreSQL, OAuth, Cloudflare
PostureNudge Shipped macOS product with privacy-first sensing, native system integration, tests, and downloadable releases. SwiftUI, AppKit, Swift
Operations Dispatcher Explores real-time operational state, persistence, recovery, map workflows, and property-based testing. React, TypeScript, Leaflet
Constraint-Aware Agent Pipeline Multi-stage research and response pipeline with rule validation, retries, fallbacks, and server-side model access. TypeScript, Gemini, Cloudflare

How I work

  • Start with the customer workflow, constraints, risks, and measurable outcome.
  • Turn ambiguous requirements into executable specifications and acceptance tests.
  • Use coding agents to accelerate implementation, review, debugging, and documentation.
  • Keep permissions, validation, confirmation, observability, and rollback in application code.
  • Evaluate real failure modes: ambiguity, stale state, unsafe actions, retries, and recovery.
  • Own the product judgment, architecture, security decisions, and final review.

Focus

Agentic workflows · enterprise AI deployment · technical customer success · solution architecture · integrations · REST APIs · OAuth and SSO · evaluation and reliability · TypeScript · Python · Swift · PostgreSQL · Cloudflare

Pinned Loading

  1. agentic-travel-planner agentic-travel-planner Public

    Privacy-safe agentic planner with grounded proposals, confirmation-bound writes, evals, and scoped undo

    TypeScript

  2. career-evidence-engine career-evidence-engine Public

    Evidence-grounded career tailoring with claim provenance, gap analysis, policy gates, and human review

    Python

  3. gtaw-roleplay-assist gtaw-roleplay-assist Public

    Constraint-aware agent pipeline with research, planning, rule validation, retries, and property-based tests

    TypeScript

  4. lssd-dispatcher lssd-dispatcher Public

    Real-time operations console with resilient state sync, persistence, recovery, maps, and property-based tests

    TypeScript

  5. posture-nudge posture-nudge Public

    Shipped privacy-first macOS wellbeing app with meeting-aware posture, blink, and eye-break reminders

    Swift

  6. receipt-tracker receipt-tracker Public

    Multi-user customer evidence inbox with AI-assisted notes, PostgreSQL, OAuth, and Cloudflare

    TypeScript