An AI-powered School & Student Information System with face-detection attendance, GenAI assistants, and computer-vision safety guardrails. Built solo from Singapore. Before this: 15 years shipping credit-risk ML for banks - most recently VP & AI/ML Lead at DBS.
I'm a technology leader and developer based in Singapore. Right now I'm building LucidLoom - an AI-native platform for schools that puts computer vision, GenAI, and a genetic-algorithm scheduler in the hands of teachers, parents, and administrators.
The throughline of the last 15 years: credit scoring, probability of default, climate risk, and stress testing for banks that move real capital. Eight years of that at DBS Singapore, leading teams of up to 12 across Singapore, India and China, deploying 88 microservices and shipping 200+ applications under MAS and Basel oversight.
My instinct is to break a hard problem into pieces small enough to reason about, then ship them - from requirements to production. Systems that hold up under audit, scale, and scrutiny.
Roots in Pune & Chennai, building from Singapore.
An AI-powered School Management System and Student Information System designed to streamline operations and meaningfully improve learning outcomes. Built ground-up: FastAPI + PostgreSQL backend, React Native Expo mobile app, Docker + Nginx with a dual-proxy security architecture, self-hosted for full data sovereignty.
Architected and deployed 88 microservices generating regional PD, LGD, EAD, ICRR and ACRR ratings across Asia - spanning AIRB, SME, climate risk, stress testing and watchlist prediction. Led a 12-person cross-regional team and drove platform migrations (Tibco → Spring Boot/Python, Oracle → MariaDB, Solaris → Linux) that cut cost by 40% and improved performance by 30%.
Built the full lifecycle of an XGBoost supervised model estimating the likelihood of a customer defaulting within twelve months. End-to-end ownership: data, features, training, and the Spring Boot services that put predictions in front of the business. Sister project: Sticky Deposit - retention-pattern analysis on the deposits book.
Solo-building the platform end-to-end: FastAPI + Postgres, React Native Expo, Dockerized self-hosted deploys. Computer vision for attendance and content safety, a genetic scheduler for timetabling, age-tuned GenAI assistants. Foundations laid for personalised learning paths and AI-generated report-card comments.
Led a 12-person cross-regional team delivering 200+ applications across credit risk modelling, regulatory reporting and AI-powered insights. Built an internal LLM assistant for risk analysts that cut compliance-report turnaround by 35%. Deployed 88 microservices in the Credit Scoring Model ecosystem.
Built CRANE - an XGBoost default-prediction model with a 12-month horizon. Contributed to the Sticky Deposit ML project. Migrated legacy Tibco BE apps to Spring Boot, cutting runtime by 30%.
Led migration of Tibco BW 5.x → 6.3 across core money-transfer systems. Built transaction business logic in Java and Tibco for Profile Service.
Managed 22 developers across Customer, Crew and Flight domains. Delivered Auto Check-In and Flight Notification System on Tibco BW/BE and Java.
SME for the WFMT Ethernet product. Built modules in Tibco, Java and J2EE. Led design and code reviews across teams.
Where the love for technology started. SCJP, OCJP, Oracle SQL certified along the way.
It started in the WFH fog of COVID - too much energy, not enough outlet. The workout habit stuck. For three years now, every morning at 4 AM, I'm paying my dues. The same discipline that ships models on deadline.