Where careful thinking meets applied intelligence
Veldra was founded on the belief that AI works best when it's built around a real understanding of your business — not adapted from a generic playbook.
Back to HomeBuilt in Kuala Lumpur, shaped by local realities
Veldra began in 2019 when a small team of data practitioners — drawn together by shared frustration with overcomplicated AI deployments — decided to take a different approach. Rather than selling large, unwieldy platforms, they focused on contained, well-scoped engagements that businesses could actually see value from.
Over the years, that focus has held. We work with mid-sized Malaysian organisations across finance, logistics, healthcare administration, and professional services. Each engagement is scoped carefully, communicated clearly, and evaluated honestly at the end.
Our name — Veldra — has no single meaning. It was chosen simply because it sounded considered, and because a name with no inherited baggage felt right for a firm trying to do things a little differently.
Mission
To make applied AI accessible and genuinely useful for Malaysian enterprises — without overpromising, overcomplicating, or overcharging.
Vision
A future where intelligent tools are embedded naturally in operations — designed by people who understood the work, not imposed from the outside.
Values
Transparency in communication, restraint in scope, rigour in execution, and respect for the humans who will ultimately work alongside what we build.
People behind the work
A small, focused team with backgrounds in machine learning, software engineering, and enterprise consulting.
Nadia Azhari
Co-Founder & Head of Modelling
Nadia leads model architecture and client data strategy. She previously spent seven years in financial analytics at a Kuala Lumpur-based investment firm before moving into AI consultancy.
Rajan Krishnamurthy
Co-Founder & Engineering Lead
Rajan oversees pipeline construction, integration, and system reliability. His background spans backend engineering and enterprise software delivery across Southeast Asia.
Lim Wei Ming
NLP & Dialogue Specialist
Wei Ming designs conversational agent flows and language understanding configurations. He brings a combined background in linguistics and applied NLP research from Universiti Malaya.
How we approach quality
Every engagement is governed by a set of working principles we've developed over years of hands-on delivery.
Defined scope before work begins
We document scope, deliverables, and evaluation criteria in writing before any development starts — so both parties know what success looks like.
Structured refinement cycles
Our services include multiple rounds of review and adjustment — outputs are measured against real-world criteria, not just technical benchmarks.
PDPA-aligned data handling
All client data is managed under confidentiality agreements and processed in accordance with Malaysia's Personal Data Protection Act 2010.
Transparent status communication
Regular check-in calls and written progress updates keep stakeholders informed throughout — with no surprises at delivery.
Documented handover
Every engagement concludes with written documentation — covering architecture decisions, operational notes, and guidance for your internal team.
Post-delivery evaluation
We schedule a structured review session after delivery to assess whether the system is performing as intended and identify any further adjustments.
Applied AI services for Malaysian organisations
Veldra operates at the intersection of data science and operational design — a space where the technical and the practical need to work together. Our engagements are structured around three core disciplines: predictive modelling for anticipating operational patterns, conversational agent development for improving digital interactions, and document intelligence for reducing manual data handling workload.
Based in Kuala Lumpur, we serve clients across the Klang Valley and broader Peninsula Malaysia. Our work is grounded in an understanding of the Malaysian business environment — including its regulatory context, multilingual digital landscape, and the diverse data maturity levels found across different sectors and organisation sizes.
We believe useful AI is quiet AI — systems that do their work without requiring constant attention or technical expertise to operate. That principle guides how we design, tune, and hand over every system we build. Our clients' teams should be able to use and maintain what we deliver; otherwise, we haven't done our job properly.
Work with a team that listens first
Every engagement starts with a conversation — reach out when you're ready to discuss what you're working with.
Contact Veldra