Requirements
AI-assisted requirements derivation for space missions
Turning mission briefs and ECSS standards into structured, traceable engineering requirements.
98%
of engineering implications captured from mission briefs and ECSS standards
Son NgoApplied AI Scientist
Agents for safety-critical work.
DiscoverSon Ngo, an applied AI scientist and PhD researcher in Melbourne, building AI agents and machine-learning systems for safety-critical, regulation-heavy work.
Applied AI Scientist and PhD researcher
- Applied AI Scientist, Starbound Space Solutions
- PhD, Monash University, National Industry PhD Program
- Generative AI for regulatory compliance
My research explores generative-AI methods that help engineers meet regulatory requirements in safety-critical domains, starting with the Australian space sector. The project is part of the National Industry PhD Program.
I build AI agents and machine-learning systems for safety-critical, regulation-heavy work, and I bring research rigour to production code.
Agents, models and systems for safety-critical work: my projects.
Voice agents
Trustlane
AI-powered EV warranty coordination that catches missing evidence before a claim is submitted.
33
behavioural scenarios in the evaluation suite
Agents
Syncnapse
Voice agents that join your meetings and turn what is said into company-owned memory.
Winner
Founder Mode Track, awarded around US$160K
Machine learning
Graph neural networks for drug discovery
Multi-task GNNs that predict pharmacokinetic (ADME) properties to speed up drug screening.
Top 13
of 370+ global participants, macro-averaged RAE 0.588
Multimodal
Jarvis
A multimodal assistant that recognises its user, then plans and delegates work through tool-calling agents.
0.904
ROC-AUC on face verification, vs 0.837 for a ResNet18 baseline
Agents, knowledge and requirements: what I bring to safety-critical teams.
AI agent orchestration
Multi-agent systems that plan, call tools and hand work to humans for review.
Knowledge representation
Turning standards, briefs and conversations into structured, traceable knowledge.
Requirements engineering
Deriving engineering requirements from regulations, with ReqIF exports engineers can trust.
Predictive machine learning
Graph neural networks, multimodal and vision models, trained and evaluated at scale.
Recognition from industry, research and university communities.
01 / 03
Awarded around US$160K at the largest buildathon in South East Asia: 2,000+ builders, 10+ enterprise partners and 500+ agentic AI projects across Asia-Pacific.
Top 13 of 370+ global participants in an AI-for-drug-discovery challenge, with multi-task graph neural networks built at CSIRO.
Member of the Golden Key International Honour Society, which recognises the top 15% of students across all disciplines.