JLJinghang LiApplied AI / Chicago
Resume ↓Email ↗

Projects / Applied AI

Applied AI, designed as decision systems.

Selected areas of work that connect models to context, constraints, evaluation, and the people responsible for the next action.

  • Agentic AI
  • Recommendation Systems
  • Semantic Knowledge
  • Adaptive ML
Selected systems

The model is one part.
The decision is the work.

01 / Agentic AI

Dependable LLM agents

AI agents designed to make progress without hiding uncertainty from the people using them.

  • Intent routing and tool orchestration
  • Grounding with semantic and operational context
  • Behavioral evaluation and observability
  • Confidence, reversibility, and human review
02 / Recommendations

Constraint-aware ranking

Recommendation systems that treat eligibility and operating conditions as first-class parts of relevance.

  • Candidate quality and contextual fit
  • Hard constraints and availability
  • Legible ranking tradeoffs
  • Outcome feedback for continual improvement
03 / Adaptive ML

Learning under change

Machine learning systems designed for shifting data, evolving conditions, and evaluation across time.

  • Drift-aware monitoring
  • Selective model updates
  • Longitudinal evaluation
  • Feedback loops connected to real outcomes
04 / Knowledge systems

Semantic context for action

Knowledge and communication layers that put the right operational context inside an AI-assisted workflow.

  • Semantic retrieval and structured context
  • Policy-aware communication
  • Traceable evidence
  • Interfaces built around a useful next action
Method

How I move from ambiguity
to a working system.

01 / Frame

Define the decision

Clarify who acts, what outcome matters, and which constraints cannot be negotiated.

02 / Evaluate

Test useful behavior

Measure quality, risk, consistency, and failure modes—not only model fluency.

03 / Learn

Ship the feedback loop

Make outcomes observable so the product, data, and model can improve together.

Research foundation

See where adaptation
started.

Explore research →