--- title: "Charlie Tang Hoong — Forward Deployed AI & Automation Engineer, Malaysia" url: https://tanghoong.com/ locale: en description: "Forward deployed AI and automation engineer in Selangor, Malaysia. 15+ years turning ambiguous business problems into production AI, automation and software systems across Malaysia, Singapore and APAC." --- # Charlie Tang Hoong **Forward Deployed AI & Automation Engineer** — Selangor, Malaysia · UTC+8 I turn ambiguous business problems into production-ready AI, automation and software systems. Fifteen years of shipping production software across AdTech, e-commerce, logistics and enterprise automation — based in Selangor, Malaysia, and working with customers across Malaysia, Singapore and the wider APAC region, remotely or on site. ## Proof - **15+ years** — Production delivery — eight roles, four employers, two ventures - **Tens of thousands req/s** — AdTech backends carrying billions of monthly ad impressions - **Sub-100 ms p95** — On selected paths, after refactoring, query tuning and multi-layer caching - **15+ pipelines** — Price monitoring, lead generation and market intelligence, built for SME clients - **3 of 3 costed honestly** — AI feasibility engagements in three unrelated domains — two declined on capital cost, one built Two of those three AI proposals did not go ahead, and that is the point: the customer got a real number early enough to say no cheaply. ## Selected Work Five deliveries, each written the same way: what the customer was actually trying to do, what was unclear, what I decided, what I built with my own hands, and how it reached production. - Case 1 — Building against a system we did not own: A multi-role booking and ordering platform went live against an enterprise client's legacy ERP, with the API contract negotiated directly with the client's own engineers and every scope change signed off before it reached engineering. - Case 2 — Taking the typing out of a document workflow that could not go offline: A hybrid OCR and LLM pipeline substantially reduced the manual data entry in a high-volume shipping-document workflow, inside an operation whose source systems had no API and could not be replaced. - Case 3 — Serving an AI assistant's traffic without holding a connection open: Re-architected an AI customer-support assistant's request path so that a conversation is queued and polled rather than held open on a synchronous connection — delivered and merged. - Case 4 — Three AI proposals, and the two I advised against: Three AI feasibility engagements in three unrelated domains, each turned into a solution design with a costed build estimate: two were correctly declined on capital cost, and the third was built. - Case 5 — The engineering floor: billions of impressions, and a marketplace that ran itself: Eleven years across two stints at the same AdTech company: four as the sole front-end engineer on a platform conversion, then a self-serve advertising marketplace built end to end, then backend services carrying tens of thousands of requests a second. ## How I Work The same seven steps whether the customer is an enterprise operations team or a two-person SME. The order matters more than the tooling: most delivery failures I have seen were decided before any code was written. - **01 Discover** — Sit with the people doing the work today and document the current process as it really runs, not as the org chart describes it. Most of what matters is not in the brief. - **02 Define** — Draw the proposed flow, walk the stakeholder through it, and get written confirmation before engineering starts. Name what is explicitly out of scope, and what is deliberately deferred to a second phase. - **03 Design** — Decide what to buy and what to build before deciding how to build it. Integrate the commodity parts, hand-build the critical path, and be honest about which constraints are real — connectivity, permissions, someone else’s release cycle, a budget that will not stretch. - **04 Build** — Take the complex, contract-defining work personally; delegate the straightforward surface. When a task has no clear direction after thirty minutes, it becomes a design discussion instead of silent rework. - **05 Deploy** — Assume the environment is not mine to change. Integrate without invading: read from what exists, checkpoint long runs so an interruption resumes instead of restarting, and always leave a path for a person to take over mid-run. - **06 Measure** — Decide up front what “working” means and how it will be observed — retry behaviour, failure visibility, cost per request, and for AI systems an evaluation set rather than an impression. - **07 Iterate** — Ship the phase, keep the second phase’s prerequisites visible, and let what production reveals reorder the backlog. ## What I Bring - **Customer Discovery & Solution Framing** — Working directly with the people who own the problem — operations staff, founders, a client’s own engineers — to turn an incomplete brief into a defined scope with a costed estimate and an agreed definition of done. I have run this from both sides: as an internal business process specialist mapping undocumented finance and logistics workflows, and as an outside engineer scoping AI projects for SMEs who had never bought one before. - **Applied AI & Workflow Automation** — AI systems that have to run inside a real operation: document extraction with a human review step before anything is committed, customer-support assistants with knowledge structure, tool integration, session handling and escalation to a person, and RPA workflows across finance and logistics tasks that used to be typed by hand. - **Solution Architecture & Enterprise Integration** — Connecting systems that were never designed to be connected — a client’s legacy ERP, source systems with no API at all, third-party commerce and payment platforms, message queues, and the manual steps in between. The recurring decision is build versus buy, and I have been on both sides of it enough to know that the interesting question is which constraint makes the bought component fail. - **Production Engineering & Reliability** — Owning a system past its release. Long-running jobs that resume from where they stopped rather than starting over. Failures that surface at a time of day when a human can still fall back to the manual process. Queue-based handling so one slow request cannot block the next customer. Submissions that cannot be double-fired by a page refresh. Idle capacity that gets released instead of held. - **Technical Delivery & Stakeholder Alignment** — I have led delivery and technical decisions rather than headcount: the final architecture and technology-selection calls after taking the team’s input, the code-review practice a team works to, and mentoring developers through the thing actually blocking them — which, in my experience, is almost always the API contract rather than the code.