[ Project ]
LOBForecasting
A public notebook project in the applied ML and forecasting cluster, useful as a case-study candidate after results and limitations are tightened.
machine-learningforecastingnotebookresearch-candidate
/ Case study
Problem, method, result, code, write-up, and status.
Problem
Forecasting projects need to make assumptions, data constraints, evaluation choices, and failure modes visible enough for others to inspect.
Method
Use notebook-based exploration to model forecasting behavior, then prepare the work for a stronger project page with setup, evaluation, and limitations.
Result
The write-up documents a strong logistic baseline, an MLP that underperforms it, a promising DeepLOB-style validation signal, and the need for a unified reproducible evaluation harness.
Status
applied ML case study candidate
/ Contact
Contact me about research, internships, consulting, or NUS AI Society collaboration.
Useful messages usually start from a specific overlap: a paper, project, team, startup problem, internship role, workshop, sponsor idea, or AI Society collaboration.
Email is the best first step. Include the context, timeline, and the kind of response that would be useful.
/ More projects
Continue inspecting the project trail.
active flagship
Computer Science Notes
A static technical hub that turns Notion notes, blog essays, search, RSS, and portfolio pages into Cloudflare Pages artifacts.
active research and toolingAgentic Coding Workflows
A public project and notes trail around controllable agent workflows, harnesses, code review, security, and maintainable AI-assisted development.