I build AI-powered systems that replace manual, repetitive work — multi-agent pipelines, RAG architectures, automated data validation, and intelligent workflow orchestration.
A RAG pipeline that ingests multiple PDF documents simultaneously and surfaces factual contradictions across them — tested on real DBS quarterly financial reports. What analysts spend hours cross-checking, the system flags in minutes, with a built-in chatbot for follow-up interrogation.
A fully automated job sourcing system that scrapes 150+ Singapore listings daily across MyCareersFuture, LinkedIn and JobStreet, scores each against a candidate profile using Claude API, routes high-relevance matches to Airtable, and delivers a curated daily digest. Distinguishes genuine AI tool roles from roles that merely mention AI in passing.
A 4-agent pipeline that takes raw data through ingestion, rule-based checks, and Claude-powered anomaly detection — catching the class of errors that Excel formulas and conditional formatting systematically miss. Outputs a structured Excel report. Replaces approximately one hour of intensive manual review per run.
An end-to-end pipeline that extracts structured ESG KPI data from PDF sustainability reports, loads it into SQLite, and surfaces it in Tableau dashboards — eliminating manual copy-paste from analyst workflows and enabling consistent, repeatable data collection at scale.
A rebuilt version of the ESG KPI extraction pipeline with MCP integration — enabling richer tool use, smarter document routing, and deeper structured output from sustainability reports.
More automation and AI tooling projects in the pipeline. This space gets updated regularly — check back or follow on GitHub.
Follow on GitHubOpen to AI automation, workflow engineering, and process automation roles in Singapore.