IT Lead · Automation, Data & AI for semiconductor manufacturing
I build the systems that take repetitive work off people's hands — RPA, AI agents, ERP integration and supply-chain analytics. Penang, Malaysia.
Eight years building automation, data and enterprise systems for high-mix manufacturing serving LAM Research and Applied Materials. I lead the IT function at Rapid Manufacturing — development, RPA, MES, ERP, analytics and security — and still write the code. The measurable part looks like this.
10+
labour hours saved every day, across five departments
6–8h → 3min
warehouse work-order printing, after automation
8,000+
PO lines analysed weekly without manual handling
5
invention disclosure awards at Keysight Technologies
20+ / 30+
production applications and internal tools shipped in three years, alongside leading the team
10,000
invoice lines processed weekly through the LHDN e-invoicing integration I built
24/7
always-on agent platform, with a live dashboard for every bot task
Team Leader of a three-person cross-functional team building an enterprise RPA and AI agent platform, now serving five departments. Verified reduction of at least 10 labour hours per day since launch, against a 15-hour target — with 24/7 always-on architecture and real-time bot monitoring. Warehouse work-order printing dropped from 6–8 hours to 2–3 minutes; engineering alternative-material extraction from 10+ hours per customer to minutes across 4,000–5,000 records.
Fully automated material allocation for the Program Management, Planning and Procurement teams. Integrated with customer portals and open-market platforms for live data on alternative materials, authorised shortages and available stock — so teams can tell instantly whether a work order is ready to build.
LHDN-compliant e-invoicing for Finance, integrating the Rapid ERP system with the government portal to automate data streaming and status sync. Processes over 10,000 invoice lines weekly.
Led recovery for the Malaysia and global teams after a major security incident: ran the module-by-module restoration programme end to end, chaired the recovery cadence with department heads, presented BCP status directly to senior management, and secured executive sign-off before returning business-critical modules to production.
Built five intelligence and security platforms, each awarded a Keysight Invention Disclosure for original, patentable engineering work.
One-stop platform capturing research grant opportunities from universities worldwide and routing them to the right sales teams, turning previously missed opportunities into pipeline. Closed millions in sales in its first year of deployment.
Centralised intelligence platform on AWS consolidating competitor product data, internal product lines, market news and R&D findings into a single view — giving Program Management full market context instead of assembling it from scattered sources. (ASP.NET Core, IIS, TensorFlow, Azure AD SSO, Python)
Deep-learning NLP pipeline monitoring competitor and market news at scale and pushing relevant items to the owning departments, shortening the lag between a market signal and the team that needed to act on it — in support of the division's first-to-market objective. (Scrapy, PyTorch, NLP, ASP.NET Core)
360° NPI competitive-analysis application giving product teams a consolidated view of new product introductions against the market, so portfolio strengths and weaknesses could be assessed against competing offerings. (ASP.NET Core, Amcharts, GemBox)
Machine-learning anomaly-scoring model that risk-classifies user login and server request behaviour, raising the application security baseline and strengthening data protection.
The company's core operations platform — maintained and extended the ERP to keep daily operations running. (Node.js, React, Ant Design, Google Cloud)
Android app for maintenance tracking with offline-first synchronisation, for use in signal-restricted areas. (Java, SQLite, MySQL, Node.js)
Cloud application ingesting sensor data for live monitoring and reporting. (Node.js, React, Google Cloud Pub/Sub, BigQuery, Data Studio)
Authentication and authorisation across Promserv's internal applications. (Node.js, React, JWT, Google Cloud)
What I use in production, grouped by what it's for.
Where most of my delivered savings come from.
Master of Data Science in progress at USM.
Every certificate below links to its public verification page.
Universiti Sains Malaysia · CGPA 3.70
Universiti Tenaga Nasional · Information Systems · CGPA 3.75
Malaysia Board of Technologists · Reg. No. PT23090096 · Active
Google Cloud Certified · ID 24491 · Valid to Sep 2027
Hands-on labs across GCP, Vertex AI, BigQuery and Kubernetes
Udacity · 2021
2022
USM · 2025 · plus E-Invoice Malaysia (2025)
Recognition across industry, invention and academia.
Also: Smarter Penang Hackathon 2018 (Top 5), Graduate Industry Excellence Award, Programming Competition 2016 & 2017.
Preprints on big-data architecture, written alongside the Master of Data Science.
Designs a unified analytics data pipeline for a multinational e-commerce platform using the Lambda architecture on AWS, with load balancing and auto-scaling to hold performance under varying transaction volumes.
Lee Chong Keat and one other · read the preprint
Compares Google Cloud Bigtable, Snowflake Storage and Apache Cassandra across financial, industrial services and transportation use cases, weighing the trade-offs of each against real-world workloads.
Lee Chong Keat · read the preprint
Open to conversations about automation, AI and IT leadership roles in manufacturing and semiconductor supply chains.