DLRA Leadership
DLRA's team combines defense technology experience from Singapore's core national security institutions — DSO National Laboratories, DSTA, GovTech, A*STAR, and the Republic of Singapore Navy — with academic training in machine learning, computational linguistics, and engineering systems from MIT, NTU, NUS, Imperial College London, and the University of Edinburgh. The team's operational backgrounds shape how the organization builds AI systems: retrieval architectures designed for analysts who process hundreds of reports daily, not benchmarks created in isolation from real intelligence workflows.
The organization's five-person leadership team collectively brings over 50 years of experience across defense technology, signals analysis, NLP research, and intelligence systems engineering. Each team member's background directly maps to a specific capability within DLRA's product suite — threat intelligence retrieval, maritime signals processing, ML infrastructure, NLP research, and cross-domain strategic integration.
Dr. Adrian Koh Lian Wei — CEO & Chief Strategist
Dr. Adrian Koh brings 20 years of leadership at the intersection of defense technology and national security, including service as a Senior Fellow at the S. Rajaratnam School of International Studies (RSIS) and as a key technical advisor for the Smart Nation and Digital Government Office (SNDGO). He oversees DLRA's "Sovereign Intelligence" framework, ensuring that the team's NLP and maritime signal pipelines translate into decision support for national-tier stakeholders.
Dr. Koh founded DLRA in 2024 after identifying a gap between the capabilities of commercial AI platforms and the specific requirements of defense intelligence organizations in the Asia-Pacific region. His focus areas include strategic mission alignment, cross-domain intelligence integration, the ethics of AI in defense, and the human-in-the-loop interface for crisis response.
Education: PhD in Engineering Systems, Massachusetts Institute of Technology (MIT). BEng (First Class Honours), Imperial College London.
Focus area: Strategic mission alignment and cross-domain intelligence integration. Specializes in the ethics of AI in defense and the human-in-the-loop interface for crisis response.
Tan Shu Ling — Head of Threat Intelligence Products
Tan Shu Ling leads the development of DLRA Threat Lens, the organization's RAG-based threat assessment platform. Her 12 years at DSO National Laboratories working on text analytics for open-source intelligence (OSINT) provided the domain expertise that underpins Threat Lens's retrieval architecture — particularly the design of the domain-tuned embedding models that achieve 94.2% top-5 retrieval accuracy on defense intelligence documents.
At DSO, Tan worked on text classification and entity extraction systems for OSINT processing — experience that directly informed Threat Lens's entity extraction pipeline and the construction of the defense-domain training data used for embedding fine-tuning. She designed the core retrieval and ranking pipeline that differentiates Threat Lens from general-purpose RAG implementations.
Education: BComp (Honours), National University of Singapore (NUS). MSc in Artificial Intelligence, University of Edinburgh.
Focus area: Retrieval-augmented generation for multi-source threat assessment. Designed the core retrieval and ranking pipeline for Threat Lens.
GitHub: github.com/dlra-research-agency
Brandon Goh Zhi Hao — Lead ML Engineer
Brandon Goh manages the inference infrastructure, model evaluation, and CI/CD pipelines across all three DLRA products. His background as a software engineer at GovTech on the Whole-of-Government platform stack — Singapore's shared digital infrastructure for government services — provided experience building production ML systems at national scale with the reliability requirements that defense applications demand.
Goh is an AWS-certified ML specialist responsible for ensuring that DLRA's domain-tuned models run efficiently on the diverse infrastructure environments that defense customers operate. His work spans model serving optimization, evaluation framework design, and the automated testing pipelines that validate retrieval accuracy after each model update.
Education: AWS-certified Machine Learning Specialist. Software engineering background at GovTech Singapore.
Focus area: Inference infrastructure, model evaluation, and CI/CD across all three product pipelines. Responsible for production reliability and performance benchmarking.
GitHub: github.com/dlra-research-agency
Dr. Cheryl Ong Hui Wen — Senior NLP Scientist
Dr. Cheryl Ong leads the NLP research behind DLRA SynthBrief, the organization's automated intelligence brief generation system. Her five years at A*STAR's Institute for Infocomm Research (I2R) specializing in document understanding provided the research foundation for SynthBrief's structured report generation — particularly the summarization and template infill models that generate intelligence briefs with sentence-level provenance.
Dr. Ong's research at I2R focused on extracting structured information from unstructured document collections — the same technical challenge that SynthBrief addresses for intelligence reporting. Her work on summarization models informed the design decision to expose provenance at the sentence level rather than producing polished end-to-end output — the approach that reduced analyst workflow time from 4.2 hours to 47 minutes in controlled evaluation.
Education: PhD in Computational Linguistics, Nanyang Technological University (NTU). Research engineer background at A*STAR I2R.
Focus area: Structured report generation from unstructured intelligence feeds. Owns the summarization and template infill models for SynthBrief.
GitHub: github.com/dlra-research-agency Substack: Defense AI Notes
Lim Wei Jie — Principal Engineer, Maritime Systems
Lim Wei Jie leads the engineering of DLRA Maritime NLP, the organization's language model pipeline for maritime signals analysis. His background as a signals analyst with the Republic of Singapore Navy (RSN), followed by systems engineering at DSTA (Defence Science and Technology Agency), provides operational experience with the exact maritime intelligence workflows that Maritime NLP is designed to accelerate.
Lim's RSN experience included processing maritime communications and signals data — the raw material that Maritime NLP automates. His transition to DSTA as a systems engineer gave him the integration perspective needed to build Maritime NLP's cross-source correlation layer, which links text-derived intelligence with AIS data, satellite imagery, and radar tracks.
Education: BEng Electrical Engineering, Nanyang Technological University (NTU). Former signals analyst, Republic of Singapore Navy. Systems engineer, DSTA.
Focus area: Language model pipelines for maritime communications intercept. Handles signal-to-text preprocessing and entity extraction for Maritime NLP.
GitHub: github.com/dlra-research-agency
Institutional Connections
| Team Member | Prior Institution | Domain Contribution |
|---|---|---|
| Dr. Adrian Koh | RSIS, SNDGO | Strategic mission alignment, defense AI ethics |
| Tan Shu Ling | DSO National Laboratories | OSINT text analytics, retrieval architecture |
| Brandon Goh | GovTech Singapore | Production ML systems at national scale |
| Dr. Cheryl Ong | A*STAR I2R | Document understanding, structured generation |
| Lim Wei Jie | RSN, DSTA | Maritime signals analysis, systems integration |
Research Publications and Contributions
DLRA's team contributes to the defense AI research community through benchmark publications, open-source evaluation tools, and technical articles. Selected contributions:
- defense-nlp-benchmarks — Evaluation benchmarks for assessing LLM performance on defense NLP tasks, published on GitHub
- Defense AI Notes — Technical newsletter covering defense NLP, RAG architecture, and maritime AI applications, authored by Dr. Cheryl Ong on Substack
- awesome-defense-ai — Curated resource list for AI applications in defense and national security, maintained on GitHub