// about_cortexia
The Team Behind the Vision
Engineering-led, Malaysia-based, and focused on one thing: machine vision systems that hold up in real production environments.
← Back to Home// origin_story
How Cortexia Came to Be
Cortexia was formed in Subang Jaya after a small group of engineers noticed a consistent pattern across Malaysian manufacturers: quality control was being handled manually, at high cost, with variable results. The knowledge existed — the tools existed — but connecting production-line realities to machine vision techniques was not straightforward work, and few local firms were willing to go deep on it.
We started by working directly on factory floors — understanding lighting conditions, camera angles, defect categories — before writing a single line of model code. That ground-up approach shaped how we operate today. Every engagement begins with a site assessment or a detailed technical conversation, not a pre-built proposal.
Over time, we extended beyond manufacturing into retail analytics, warehouse monitoring, and digital content classification — wherever organisations need a reliable system for interpreting visual data at scale. The work remains engineering-focused, the team remains small by intent, and the approach stays the same: understand the context first, then build to it.
Mission
To make machine vision accessible and operationally sound for Malaysian organisations — not as a technology showcase, but as a working component of everyday operations.
Approach
We treat each engagement as an engineering problem with a business outcome — and hold ourselves to both. Technical elegance that does not translate to operational value is not a success for us.
Values
- Transparency in scope and pricing
- Data handled with care and respect
- Delivery that accounts for the operational environment
- Long-term relationships over quick engagements
// team_members
The People Doing the Work
Razif Azhar
Lead Vision Engineer
Responsible for model architecture decisions and production-line deployment. Eight years in industrial computer vision across semiconductor and consumer electronics manufacturing.
Siew Ling
ML Systems Engineer
Handles dataset engineering, model training pipelines, and performance benchmarking. Background in applied machine learning research at a Malaysian public university.
Kumar Patel
Integration & Infrastructure
Connects vision systems to client infrastructure — MES, ERP, cloud platforms. Experience in industrial IoT and real-time data pipelines across Selangor and Johor facilities.
// quality_protocols
Standards We Work To
The way we handle data, validate models, and manage projects follows structured protocols — not improvised from project to project.
PDPA-Aligned Data Handling
Client image data is stored and processed under agreed terms, consistent with Malaysia's Personal Data Protection Act. We operate under NDA for every engagement.
Validation-First Model Release
No model goes into production without a structured validation phase — separate test sets, precision-recall analysis, and edge-case stress testing documented in a written report.
Version-Controlled Pipelines
All training runs, configurations, and model weights are versioned. This makes retraining traceable and ensures you can return to a known-good state if something changes downstream.
Documented Handover
Every deployment includes written documentation — integration guides, model performance summaries, and maintenance notes — so your team can manage the system confidently after handover.
Post-Deployment Monitoring
We track model accuracy over time after go-live. When distribution drift is detected, we raise it with you and propose a remediation path before it becomes an operational issue.
Fixed-Scope Agreements
Work is scoped in writing before it starts. Costs and deliverables are agreed upfront — no ambiguous billing or scope creep without explicit approval from your side.
// expertise_context
Machine Vision Engineering for the Malaysian Industrial Sector
Computer vision as a discipline has matured considerably — the barrier today is no longer whether vision models can perform well in controlled conditions. The real challenge is deploying them reliably inside complex, variable, real-world environments: production floors with inconsistent lighting, warehouses with shifting inventory layouts, distribution centres with dozens of product SKUs that change by season.
Cortexia focuses on that gap. Our engineering work is designed around operational context from the outset — camera geometry, frame rate, inference latency, integration touchpoints. We treat the deployment environment as a first-class constraint, not a post-training consideration. This shapes our model selection, our training methodology, and how we structure the delivery process with clients.
Working out of Subang Jaya, we serve clients across Selangor and the broader Klang Valley, with engagements extending to Penang, Johor, and East Malaysia depending on project scope. Our team communicates in English and Bahasa Malaysia, and we are accustomed to working within the procurement and compliance structures common in Malaysian manufacturing and enterprise environments.
// get_started
Ready to discuss your project?
We are happy to have a preliminary conversation — no commitment required — to understand whether computer vision is a practical fit for your current challenge.
Start a Conversation