Cameron Loh

AI-driven multi-agent financial assistant integrating Stock Agent (market research & trading insights) and Invoice Agent (document parsing, reconciliation & AR automation). Built a LangChain + LangGraph system using OpenAI APIs and AlphaVantage for real-time financial analytics, supporting up to 1,000 stock queries/day. Implemented a RAG pipeline with FAISS and OpenAI embeddings to improve document retrieval accuracy in hybrid ranking tasks (~10–15%). Developed an Invoice Agent capable of evaluating 1,000+ accounting/compliance rules, analyzing invoices, POs, and GRNs to detect inconsistencies in under 15 seconds. Implemented agent routing, tool orchestration, and Langfuse observability to achieve ~99% workflow reliability and enable stateful reasoning for financial automation. Integrated browser automation via Playwright and backend orchestration for stateful reasoning, traffic analysis, and autonomous exploration during simulated attacks. Developed automated vulnerability assessmen

Designed a scalable backend serving ~5,000 daily API requests with efficient caching and asynchronous query handling to achieve sub-100 ms response times. Built 10+ core modules (reservations, customer analytics, and order tracking) with ~98% uptime under a simulated load of 100 concurrent users. Adopted Agile sprints, Postman API testing, and CI/CD pipelines. Earned Artemis—Highest Distinction among ~600 teams in US Orbital for a production-ready CRM platform for restaurant operators.

Built a Next.js frontend with Zapier (iPaaS), OpenAI, and Google APIs to process up to 500 documents/day, enabling dynamic rule-based summarization. Integrated AWS WebSocket services for real-time updates, and implemented automated testing with Jest. Set up CI/CD pipelines, leveraging Vercel and OpenAI APIs for scalable processing.