Meraldo Antonio

My core quantitative work is in time series forecasting and nowcasting for macroeconomic and financial market variables — inflation, GDP growth, interest rates. I work with classical statistical methods (ARIMA, VAR), decomposable method (Prophet) and neural approaches, always with explainability as an important consideration. Beyond training, I deploy these models end-to-end: interactive dashboards in React/D3, plotly or Tableau; APIs in FastAPI, running on AWS or internal DBS platform. I also contribute to sktime and skpro, open-source Python libraries for probabilistic forecasting, which keeps me close to the research frontier.

PyTorch • scikit-learn • Tableau • plotly • React/NextJS • D3 • AWS • GCP

My work involves mapping manual workflows and designing AI-augmented alternatives that keep humans in control.

Since early 2026, my focus has shifted toward implementing agentic AI across DBS Finance: moving the department from manual workflows to AI-augmented processes. This means less model-building and more change work: mapping and understanding existing processes, designing human-in-the-loop automation, getting buy-in from stakeholders, and building the tooling that makes new workflows actually stick. I learned that organizational problems are just as hard as if not more than technical ones!

Process Design • BPMN • Claude Cowork • Microsoft Copilot • Stakeholder Management

Designing RAG systems is part of what I do. This diagram shows a simple RAG system that uses both embedding-based and BM-25-based retrieval.

I build agentic systems using tools like LangChain, LangGraph, Google SDK and Claude's API. This often involves connecting LLMs to internal data through RAG pipelines and MCP/connectors. For faster iteration, I reach often for n8n/Cowork. The interesting problems are almost never the LLM itself: they're data quality, system design, and deciding when not to use an agent at all!

Claude API • LangChain • LangGraph • n8n • Claude Cowork • MCP • RAG