Advisory Work Grounded in Practical Compute Knowledge
Tensaro was formed to fill a specific gap: mid-sized organisations in Malaysia that need structured guidance on AI accelerator infrastructure, without being steered by vendor incentives.
Back to HomeHow Tensaro Came Together
Tensaro was founded in Kuala Lumpur in 2021 by a small group of engineers and operations specialists who had spent years watching organisations make costly accelerator hardware decisions — not because the organisations lacked intelligence, but because they lacked a neutral party to help them think through the problem early enough.
The pattern was familiar: a leadership team decides to run AI workloads, an engineering lead starts sizing hardware, a procurement cycle begins, and somewhere in that chain a foundational question goes unasked — "Is our data and team actually ready for this?" By the time advisory input arrives, budget is committed and choices are harder to revisit.
We designed our three engagements to interrupt that pattern at the right moments. The Readiness Assessment is meant to happen before procurement conversations start. The Workload Planning Workshop fits neatly into early technical scoping. The Infrastructure Retainer supports teams already in motion who need a steady outside perspective.
What We Are Here to Do
Our mission is to help mid-sized organisations in Malaysia make well-reasoned decisions about AI compute infrastructure — decisions that hold up over time, that the board can understand, and that the engineering team can implement without having to discard six months of work later.
Clarity over complexity
We write plain-language deliverables. If something cannot be explained simply, we have not understood it well enough yet.
Vendor neutrality as a principle
We hold no reseller agreements. Our recommendations follow your workload requirements and procurement constraints, nothing else.
Respect for the client's pace
We do not create urgency. We work at the pace that lets your team absorb findings and act on them thoughtfully.
The People Behind the Engagements
Our advisors bring direct experience from infrastructure engineering, operations planning, and enterprise technology strategy.
Ahmad Razif bin Kamal
Principal Advisor, Compute Strategy
Ahmad spent twelve years in infrastructure engineering roles across Malaysian telecommunications and financial services before moving into advisory work. He leads readiness assessments and retainer engagements.
Lim Yi Xuan
Senior Advisor, ML Infrastructure
Yi Xuan focuses on GPU workload planning, having previously led ML platform engineering at two Kuala Lumpur–based technology companies. She facilitates all workshop engagements.
Nurul Syafiqah Mohd Yusof
Advisory Lead, Operations & Procurement
Syafiqah brings a background in enterprise procurement and vendor negotiation across the ASEAN region. She manages retainer engagements and procurement-timing advisory for scaling teams.
How We Maintain Quality Across Engagements
Our protocols are practical rather than ceremonial. They exist to make our work consistent, verifiable, and genuinely useful.
Defined Engagement Scope
Every engagement begins with a written scope document. Both parties sign before work starts. Changes require a documented amendment.
Confidentiality by Default
All client information — architecture data, procurement plans, team assessments — is treated as confidential. We operate under NDA for every engagement.
Vendor-Neutral Process
Our advisory outputs are reviewed internally against a neutrality checklist. No output should favour a specific hardware vendor without explicit, documented reasoning tied to the client's workload profile.
Written Deliverables Standard
Every engagement produces a written output. Verbal sessions are summarised in a follow-up note within five working days. Nothing is considered delivered until it is documented.
Internal Peer Review
Readiness scorecards and infrastructure roadmaps are reviewed by a second advisor before delivery. We do not send first drafts to clients.
Structured Feedback Loop
At the close of every engagement we ask for structured feedback. Responses inform how we adjust our methods for subsequent engagements.
AI Compute Advisory for the Malaysian Market
Tensaro operates at a specific intersection: the point where an organisation's leadership has decided that AI workloads are worth pursuing, but the engineering and operations teams have not yet committed to a hardware or infrastructure approach. That window — typically three to six months before a first procurement — is where our advisory work is most useful.
Accelerator-based AI compute has developed into a complex area. The choices organisations face involve not just GPU specifications but memory bandwidth, interconnect requirements, software stack compatibility, and cooling considerations. For teams without dedicated AI infrastructure specialists, these decisions carry real risk — not because the technology is too difficult, but because the framing questions are easy to miss.
Our Readiness Assessment addresses this directly. Rather than beginning with hardware, we begin with data: what workloads your team expects to run, what data assets those workloads require, and whether the current team composition can support an accelerator deployment. The output is a plain-language report that gives leadership and engineering a shared starting point.
For teams further along, the GPU Workload Planning Workshop moves from assessment into practical sizing. We use vendor-neutral frameworks to walk through memory requirements, batch processing approaches, and throughput modelling, leaving the team with templates they can reuse as their workload profile changes.
The Compute Infrastructure Retainer is designed for the organisations that have moved past initial planning and are now managing a multi-stage rollout. Regular advisory calls, a shared decision log, and roadmap documentation give the in-house team an external reference point without adding permanent headcount.
Talk to Tensaro About Where You Are in the Process
We keep initial conversations short and specific. Tell us your organisation's size, what AI workloads you are considering, and where the planning process currently sits.
Start a Conversation