Skip to content

Featured in the BEAMP Programme by Singapore's Building and Construction Authority (BCA), JTC Corporation and EnterpriseSG:

Entrepreneurial Experience

CTO & Co-Founder
FRONTLINE INDUSTRIAL SOFTWARE PTE. LTD (UEN: 202138635W) | Singapore
Nov 2021 - Present

  • Participated in Entrepreneur First's startup incubation program (2021)
  • Received initial investment and incorporated the company in Singapore in November 2021
  • Was hired by the company as CTO, holding EntrePass initially and then Employment Pass
  • Received USD $700K seed investment in early 2023 from Singapore's Cocoon Capital and Plug and Play
  • Served as CTO until present day; the company is now being acquired.

Technology & Product

Schedule Optimization

Core Innovation: Developed novel differentiable optimization algorithm for project scheduling

  • Designed and implemented from scratch a differentiable optimization approach for project scheduling problems
  • Transformed traditionally non-differentiable schedule optimization into a differentiable problem using computational graphs, making project properties (duration, cost) differentiable with respect to optimization parameters
  • Applied gradient descent and backpropagation inspired by neural network training to large-scale project optimization
  • Engineered mathematical solutions for handling discrete scenarios in project scheduling
  • This approach overcomes limitations of traditional derivative-free optimization methods when handling large parameter spaces

Maintenance Scheduling

  • Modeled tank maintenance scheduling as Mixed Integer Linear Programming problem
  • Used existing optimization libraries with Branch and Cut algorithm
  • Implemented scenario analysis with random delay simulations
  • Built interactive visualization for Gantt charts and maintenance distribution

Overrun Forecasting

  • Used OpenAI's embedding models for NLP vectorization of task names
  • Implemented HDBSCAN clustering algorithm for grouping similar tasks
  • Applied cosine similarity for task assignment to clustered groups
  • Built statistical forecasting using duration overrun ratios

AI-Assisted Planning

  • Used OpenAI's GPT models for task generation (structured output)

Technical Note: The schedule optimization engine represents original algorithm development, while other systems leverage existing libraries and algorithms with custom integration and application to domain-specific problems.

Product Website

Selected Screenshots