Our Approach to Thoughtful AI
We build AI systems that acknowledge what they don't know, cite their sources, and support human decision-making rather than replacing it.
Return HomeWho We Are
Veridis was established in early 2023 by a group of engineers and domain specialists who had grown concerned about the widening gap between AI capabilities and organisational readiness. We observed many implementations failing not due to technical limitations, but because systems were deployed without adequate consideration of their failure modes, interpretability requirements, or integration with existing decision processes.
Our founding principle holds that artificial intelligence should amplify human judgment rather than obscure it. This means building systems that clearly communicate their confidence levels, provide traceable reasoning paths, and gracefully handle questions they cannot answer. We focus on domains where we can combine technical capability with substantive expertise in the problem space.
Based in Singapore, we work primarily with organisations in Southeast Asia seeking to implement AI in sustainability measurement, knowledge management, and operational risk contexts. Our team includes specialists in machine learning engineering, sustainability science, information architecture, and organisational change management.
We believe the most valuable AI applications emerge from deep engagement with specific problems rather than pursuit of general capabilities. Our work emphasises careful scoping, realistic expectations, and sustainable integration with existing workflows and decision structures.
Our Team
Maya Wong
Technical Lead
Maya leads our machine learning development with a background in natural language processing and a particular focus on uncertainty quantification in AI systems.
Arjun Ramesh
Sustainability Domain Expert
Arjun brings deep expertise in environmental data analysis and ESG frameworks, ensuring our sustainability analytics tools produce meaningful insights.
Li Chen
Implementation Specialist
Li manages client engagements and system integration, with extensive experience helping organisations adopt complex analytical tools thoughtfully.
Our Quality Standards
Data Protection
We comply with Singapore's Personal Data Protection Act and implement appropriate technical safeguards for sensitive information. Training data remains under client control with clear access protocols.
Development Methodology
Our development process emphasises reproducibility, version control, and systematic testing. We document assumptions, limitations, and expected failure modes for all systems we deliver.
Client Collaboration
We view clients as domain experts and structure engagements to incorporate their knowledge throughout development. Regular review points ensure alignment with actual needs.
Documentation Standards
We provide comprehensive documentation including model architecture, training procedures, validation results, and maintenance requirements. Technical writing emphasises clarity over comprehensiveness.
Performance Monitoring
Systems include built-in monitoring capabilities to track accuracy, response times, and error patterns. We establish baseline metrics and degradation thresholds during deployment.
Knowledge Transfer
Each engagement includes training appropriate to whoever will maintain the system. We encourage organisations to develop internal capabilities for routine maintenance and updates.
Our Values
Epistemic Humility
We design systems that clearly communicate uncertainty and acknowledge limitations. Overstating capability does more harm than understating it. Our solutions include mechanisms for users to understand confidence levels and identify edge cases where human review is warranted.
Domain Expertise Integration
Technical capability alone produces brittle solutions. We invest time understanding the problem domain, consulting with subject matter experts, and incorporating established frameworks rather than reinventing them. This approach yields tools that align with how practitioners actually think about their work.
Practical Preparedness
AI systems will fail in unexpected ways. Rather than treating this as a theoretical possibility, we help organisations develop concrete response procedures, test them through exercises, and refine them based on operational experience. Preparedness requires planning specific actions, not just acknowledging risks.
Sustainable Integration
Technology that cannot be maintained becomes technical debt. We prioritise approaches that work within existing infrastructure, use widely-supported tools, and can be understood by teams beyond the original developers. Short-term sophistication that creates long-term dependency serves no one well.
Work With Us
If you're seeking AI development partners who prioritise transparency and practical outcomes over theoretical capabilities, we'd welcome a conversation about your needs.
Get in Touch