◈ Tessera
Tessera team at work

About Tessera

An AI firm built on
measured thinking

We have spent seven years working alongside Malaysian organisations to bring structured AI capability into their operations — at a pace that works for each team.

Back to Home

How Tessera came to be

Tessera was established in Kuala Lumpur in 2018 by a group of data engineers and machine learning practitioners who had spent years working inside large technology organisations. What they observed, repeatedly, was that the companies most in need of analytical capability were precisely those least equipped to bring it in-house.

The idea behind Tessera was straightforward: create a small, focused firm that could take on well-defined AI work for organisations across Malaysia — without the overhead of a large consultancy and without the abstraction of an off-the-shelf product.

The name comes from the individual tile pieces used in mosaic work. Each piece is modest on its own. Together, they create something coherent. We think of each engagement the same way — a distinct contribution that fits into the larger picture of how an organisation uses data.

2018

Year founded

Established in Kuala Lumpur by practitioners with backgrounds in data engineering and applied machine learning.

80+

Engagements completed

Across natural language processing, recommendation systems, and data infrastructure projects.

3

Core service areas

Natural Language Analytics, Recommendation Engine Design, and Data Pipeline Optimisation.

What guides our work

Our mission is to make applied AI accessible to Malaysian organisations through engagements that are honestly scoped, clearly communicated, and designed with the team's long-term capability in mind — not just the immediate deliverable.

Clarity over complexity

We present findings and progress in language that matches the audience. Technical depth is available when needed; it is not the default mode.

Scope before start

Work does not begin until the engagement boundaries are agreed. This protects both the client's budget and our team's ability to deliver well.

Built to be extended

Every system we build includes documentation sufficient for your team to maintain, adjust, or expand it. Our goal is your independence, not dependence.

Who you work with

AH

Ahmad Hafizi

Co-Founder · Data Engineering Lead

Leads data pipeline and infrastructure engagements. Background in large-scale ETL systems within Malaysian financial services.

SR

Siti Raihanah

Co-Founder · NLP & Analytics Lead

Specialises in text analysis and sentiment modelling. Developed taxonomy systems for regional e-commerce and government survey data.

LC

Lim Chee Wei

Recommendation Systems Engineer

Builds personalisation frameworks for content and commerce platforms. Experienced in collaborative filtering and A/B testing infrastructure.

How we approach quality

Data Privacy

All engagements operate within a documented data handling agreement. We do not retain client data beyond the project period. Processing is limited to what is necessary for the stated scope.

Documented Deliverables

Every project concludes with written documentation: system architecture, configuration details, operating procedures, and recommendations for future development.

Testing & Validation

Systems are tested against defined criteria before handover. For recommendation and analytics work, this includes performance benchmarks agreed at the start of the project.

Regular Communication

Progress updates are shared on a regular cadence throughout each engagement. Issues or scope changes are raised promptly rather than at project end.

Knowledge Transfer

Handover sessions are included in every engagement. Your team receives the context needed to operate what we build without requiring continued access to Tessera.

Ethical AI Practice

We consider potential bias in datasets and model outputs as part of standard review. Findings are communicated clearly and incorporated into system design decisions.

Applied AI in the Malaysian market

Tessera operates within a specific corner of the AI services landscape — one focused on analytical and engineering work rather than off-the-shelf products. The organisations we work with typically have collected data over time and want to draw more structured conclusions from it, or they want to offer their users more personalised experiences without building a data science function from scratch.

Our work in natural language processing covers the full cycle from raw text data to interpretable outputs. This includes designing the categories and structure needed to make sense of written feedback, building models that identify sentiment and intent, and creating dashboards that present those findings in a form useful to non-technical stakeholders.

In recommendation engine work, we focus on understanding the relationship between user behaviour and the content or products an organisation offers. The outcome is a system that surfaces relevant options at the right moment — improving engagement without requiring manual curation at scale.

Data pipeline work underpins both of these areas. Before any analysis or recommendation system can function well, the data feeding it needs to be reliable, consistently structured, and efficiently processed. We approach this as foundational work — methodical rather than complex, and necessary before more visible AI capabilities can deliver value.

Work with a team that keeps things clear

Have a project in mind? We are happy to discuss scope and approach before any commitment is made.

Start the Conversation