Singapore-based AI engineer | AI platforms | evaluation systems
Wes Lee
Builds backend-first AI platforms, multimodal evaluation systems, and operator-facing decision tools.
Currently AI Engineer and ASEAN Education Program Director at Elice. This site focuses on inspectable systems and selected private-code case studies across retrieval, evaluation, finance, forecasting, and decision support.
- AI platforms
- Evaluation systems
- Retrieval and orchestration
- Operator-facing products
Flagship case studies
Start with four flagship case studies that best represent the portfolio.
This set is intentionally small: the clearest signal for platform architecture, evaluation rigor, and operator-facing product execution.
Edtech / AI platform
Creator AI Platform
A backend-first orchestration platform for learning-asset generation with staged discovery, retrieval, validation, and human review built into the product itself.
Document intelligence / service architecture
Intelligent Content Analyzer
A modular document intelligence platform that splits upload, retrieval, generation, and evaluation into services rather than one monolithic demo app.
Evaluation systems / multimodal
ArtifactBench AI Evaluation Workbench
A multimodal evaluation workbench for AI-generated artifacts, decks, PDFs, and code bundles, combining deterministic metrics, retrieval, provenance checks, and judge-assisted scoring.
Finance / foundation models
AI Portfolio Advisory System
A robo-advisor platform that applies TabPFN to investor risk assessment and pairs it with objective-aware portfolio optimization.
Current work
Current work centers on orchestration, evaluation, and operator review.
AI platforms
Backend-first systems that connect retrieval, generation, validation, and human-in-the-loop review.
Evaluation and observability
Benchmarking, evidence checks, regression coverage, tracing, and structured scoring for model behavior.
Product interfaces
FastAPI, Streamlit, and operator-facing interface patterns that give users usable surfaces instead of opaque models.
Decision domains
Education, workforce intelligence, finance, public health, and business analytics where systems need to drive action.
Recent writing
Writing behind the systems.
Technical notes on implementation choices, evaluation logic, and what changed between experiment and usable product.
May 2, 2026
Building Longevity Lab for Health Scenario Modeling
Why health-risk interfaces need visible data provenance, model cards, runtime mode, and causal-analysis boundaries.
Read article
Mar 24, 2026
Designing Creator AI as a Backend-First Platform
Why AI content generation becomes more valuable when orchestration, discovery, validation, and review are part of the product.
Read article
Mar 24, 2026
Building ArtifactBench for AI-Generated Artifacts
How deterministic metrics, benchmark retrieval, provenance checks, and judge-assisted scoring support broader artifact evaluation.
Read articleNext move
Need an engineer who can connect model quality with product execution?
Best fit is backend-first AI platforms, evaluation-heavy systems, and decision-support products moving beyond prototype stage.
