Building Recommendation Systems That Actually Help
We're a Subang Jaya team focused on helping Malaysian businesses implement personalization in ways that genuinely serve their users.
Back to HomeOur Story
curatiixxs was founded in early 2022 by a group of data scientists and engineers who had spent years building recommendation systems for companies across Southeast Asia. We saw a consistent pattern: businesses wanted personalization capabilities, but many implementations either felt superficial or created unintended consequences like filter bubbles that limited user discovery rather than enhancing it.
We established our practice in Subang Jaya with a different approach. Instead of treating recommendation systems as a feature to check off, we work with businesses to understand what they're actually trying to achieve for their users. Sometimes that means a full recommendation engine. Other times it means simpler personalization approaches or improving what's already there. The goal is always to create systems that help users find what they'll value while supporting the business's objectives.
Our team brings together expertise in machine learning, data engineering, and user experience design. We've worked with e-commerce platforms, content publishers, educational providers, and marketplace applications across Malaysia and neighboring countries. Each project has taught us something about the nuances of recommendation systems — from handling sparse data to balancing diversity with relevance to explaining recommendations in ways users can understand.
What guides our work is a belief that personalization should expand possibilities for users, not narrow them. We design systems with transparency, measure both intended and unintended effects, and build in controls that let businesses adjust recommendations as their understanding of users evolves. The technical implementation matters, but so does the thoughtfulness behind it.
Meet Our Team
A group of specialists who bring different perspectives to recommendation system design and implementation.
Darren Raj
Lead Data Scientist
Specializes in collaborative filtering algorithms and building evaluation frameworks that capture both accuracy and diversity in recommendations.
Lina Wong
Machine Learning Engineer
Focuses on data pipeline architecture and implementing hybrid recommendation approaches that combine multiple signal types effectively.
Ahmad Ibrahim
Solutions Architect
Works with clients to translate business requirements into technical specifications and ensures recommendation systems integrate smoothly with existing platforms.
How We Approach Quality
Our standards for building recommendation systems that perform reliably and serve users well.
Comprehensive Evaluation
We measure recommendation quality through multiple lenses including accuracy, diversity, coverage, and business impact. Offline metrics are validated through live A/B testing.
Privacy Considerations
User interaction data is handled according to data protection requirements. We design systems that can deliver personalization while respecting privacy constraints.
Clean Implementation
Code is documented, tested, and structured for maintainability. We provide technical documentation that helps your team understand and operate the system.
Bias Awareness
We actively look for potential biases in recommendation outputs and implement monitoring to catch unintended patterns that could disadvantage certain users or content.
Continuous Improvement
Recommendation quality degrades over time as user preferences and content catalogues change. We design systems with mechanisms for ongoing model updates and performance monitoring.
Clear Communication
Technical complexity shouldn't prevent understanding. We explain how systems work, what trade-offs exist, and what outcomes you can reasonably expect.
What Guides Our Work
We believe recommendation systems should serve both users and businesses, not just optimize for short-term engagement metrics. This means designing systems that help users discover valuable content or products they might not have found otherwise, while supporting business objectives like increased sales or longer sessions. When these goals align, recommendations work well. When they conflict, we make the trade-offs explicit.
Transparency matters in recommendation work. Users deserve to understand why they're seeing certain suggestions. Businesses need to know how their systems make decisions and what factors influence recommendations. We build explainability into our systems where it adds value and document the logic behind algorithmic choices.
Technical sophistication serves the goal, it doesn't define it. Sometimes a simpler approach delivers better results than a complex one. We select methods based on what fits your data characteristics, business context, and operational capabilities. The most advanced algorithm isn't always the right choice.
Recommendation systems exist within larger product experiences. We consider how personalization affects the overall user journey, work with your existing technology stack, and design implementations that your team can operate and improve over time. The best recommendation engine is one that actually gets used and maintained.
Let's Explore What's Possible
If you're considering recommendation capabilities for your platform, we'd be happy to discuss your situation and share our perspective on what might work well.
Get in Touch