What clients say about working with us
A selection of feedback from research engagements across Singapore, alongside case studies that describe the work and what was achieved.
Back to HomeFrom recent research engagements
"The literature review we commissioned gave us a genuinely current picture of NLP approaches in our domain. The annotated bibliography alone has been referenced by our internal team multiple times since delivery. It was structured with us as the audience, not a generic reader."
Li Ting Ng
Head of AI Strategy, fintech firmFebruary 2026
"We needed someone to help us frame a research question that had been sitting unresolved for months. The methodology advisory sessions were specific and structured — we came out with a document we could actually use. The session notes were detailed and honest about our constraints."
Roshan Kumar
Research Lead, government tech agencyJanuary 2026
"What I valued most was that they told us early on what the data environment made difficult. A lot of firms would have pushed ahead and delivered something that looked complete but was weaker for it. Novarc Labs flagged the issue in week two and we redesigned the scope together."
Hui Wen Loh
VP Research, regional healthcare groupFebruary 2026
"The technology landscape report informed our AI procurement decision in a way that would have taken our internal team months to replicate. The executive summary was genuinely readable for our board — something harder to achieve than it sounds."
Adrian Yeo
CTO, Singapore logistics companyJanuary 2026
"We worked with Novarc Labs on a research partnership covering a genuinely novel ML evaluation problem. The methodological documentation they produced has since been used as a reference point for two internal projects. It holds up well under scrutiny."
Selina Mah
Director of R&D, enterprise software firmDecember 2025
"The scoping conversation itself was useful — it helped us identify that our original research question was actually two distinct questions, one more tractable than the other. We narrowed the scope and got a much stronger result. That kind of thinking is what you pay for."
Thanh Nguyen
Principal Scientist, research instituteFebruary 2026
Three engagements, described in detail
AI document classification in a regulated sector
CHALLENGE
A Singapore financial services organisation needed to understand current AI-based document classification approaches before a technology investment. Internal knowledge was limited and the procurement timeline was tight.
APPROACH
Structured literature review covering transformer-based classification, performance in low-resource settings, and limitations in regulated data environments. Delivered in four weeks with an annotated bibliography of 38 sources.
OUTCOME
The report directly shaped the vendor evaluation framework used in the subsequent RFP. Three of five evaluation criteria were drawn from the limitations section. Procurement was completed within the original timeline.
Evaluation methodology for an AI-assisted triage system
CHALLENGE
A healthcare technology team needed a rigorous evaluation framework before clinical deployment of their AI triage tool. Standard metrics were insufficient, and internal AI evaluation expertise was limited.
APPROACH
Eight-week research partnership covering evaluation literature for high-stakes AI, development of a context-specific framework, and a pilot evaluation plan aligned to deployment constraints and available data.
OUTCOME
The framework was adopted as the internal standard for AI system assessment across the organisation. The pilot evaluation identified one significant performance gap before clinical deployment. The report was reviewed by the ethics board.
Research design for an internal NLP benchmarking study
CHALLENGE
A government AI team wanted to benchmark NLP models against their specific use cases but were uncertain how to design the study to produce valid, comparable results.
APPROACH
Three advisory sessions covering research question formulation, benchmark dataset design, evaluation metric selection, and controls for validity threats. Written notes after each session, followed by a methodology document.
OUTCOME
The team completed the benchmarking study using the methodology document as their primary reference. The study informed a model selection decision and was included as supporting documentation in a budget proposal for further AI capability investment.
What the work adds up to
Ready to discuss your research question?
The conversation is informal and without commitment. We will give you an honest view of whether and how we can help.
Start the ConversationSingapore 117610
Mon–Fri: 9:00 am–6:00 pm SGT
Sat: 10:00 am–1:00 pm SGT