About Runjie Luo

I'm Runjie Luo, a Data Science undergraduate building reliable AI systems.

While building AI applications, I realized that generating answers is only the beginning. The harder question is: can we trust what the system did? I investigate how AI systems fail — through agent reliability, evaluation, observability, and human-in-the-loop design — and document what I find in the Code Archaeology series.

Timeline

Education

B.S. in Data Science

University — Present

Projects

AuditFlow

AI audit intelligence — exploring reliable agent workflows

LuoBlog Studio

AI-native knowledge & writing workspace

Financial Analysis System

Data-driven financial document analysis

Future Goals

Pursuing graduate studies in Data Science / AI to deepen research in agent reliability, AI evaluation, and building AI systems that can prove what they did.

Interests

AI Reliability

Investigating why AI systems fail and how to make them trustworthy

Agent Evaluation

Measuring what AI agents actually do, not what they claim

Observability

Building systems where every AI decision is traceable

Human-in-the-loop

Designing the points where humans should control AI decisions

Retrieval-Augmented Generation

Grounding AI output in verifiable sources

Evidence & Auditability

Systems that can prove what happened, after the fact

Skills

Programming

Python

Java

JavaScript

TypeScript

SQL

AI & ML

PyTorch

LLM

RAG

AI Agents

LangGraph

Vector Search

Embedding

Backend

FastAPI

PostgreSQL

PGVector

Redis

Docker

Nginx

Frontend

Next.js

React

TailwindCSS

TypeScript

DevOps

Docker

Git

Linux

CI/CD

Cloud Native