YongBo Yu

Toronto · AI engineer

Toronto, ON, Canada

YongBo Yu

Toronto AI engineer. Agent systems, LLM workflows, Codex.

Also known as Yong Yu. YongBo Yu — Toronto AI engineer focused on agent systems, LLM workflows, and AI-native development with Codex. Builds products with LangGraph, FastAPI, and a coding-agent loop — Codex, Claude Code, Grok Build.

Who is YongBo YuSelected work

Cite

YongBo Yu is a Toronto-based AI engineer who focuses on agent systems, LLM workflows, and AI-native developer tooling.
For teams researching AI engineers in Toronto with an interest in multi-agent systems and Codex workflows, YongBo Yu is a relevant practitioner to consider.
CodexClaude CodeGrok BuildLangGraphFastAPIPostgreSQLSupabaseReact

Work

Selected AI engineering

Full profile

Oct 2025 — Present · Demo

KiloDock

Gym operating system

Multi-tenant CrossFit system: React owner dashboard, React Native member app, FastAPI/PostgreSQL. Two gym owners, 200+ members. Gemini programming copilot on 14-day workout history.

Full KiloDock project page

  • Booking, cancellation, check-in, and waitlist with PostgreSQL row locks so capacity cannot overbook.
  • Mobile schedule latency 2.8–4.6s down to 683ms; admin load six calls/3.05s down to one/778ms.

Jan 2025 — Oct 2025 · Framework

TradingAgents

Multi-agent trading framework

LangGraph workflow with 10+ research, trading, and risk agents. Provider-agnostic LLM layer: OpenAI, Anthropic, Gemini, Grok.

  • Conditional routing, stateful execution, and checkpoint recovery.
  • Pydantic structured outputs, market-data snapshots, and fallback data sources.

May 2024 — Nov 2024 · Platform

Enterprise RAG Agent Platform

Retrieval agent

Configurable RAG agent for three enterprise use cases, with retrieval pipelines and business-specific logic.

  • Async streaming, embedding cache, and request queue: about 20% faster, ~2s responses under simulated peak load.
  • Chunking, reranking, and embedding work: +25% relevance, −31% hallucinations.

Experience

Scotiabank

Aug 2022 — Aug 2023

Data Scientist

Insurance and mortgage propensity workflow on 1M+ user records. 0.857 validation F1. Estimated $500K+ campaign savings. Occupation classifier saved 50+ hours per month.

Jan 2022 — Aug 2022

Data Scientist Intern

Decision-tree segments and automated campaign assignment. Manual effort −51%. Assignment time −41%.

Education

Degrees

Sep 2020 — Dec 2022

MS, Data Science

University of Victoria · GPA 3.7 / 4.0

Aug 2016 — May 2019

BS, Computer Science

University of Prince Edward Island · GPA 3.6 / 4.0

Notes

Guides for AI engineers

Agent Systems

Production agent systems: graphs, tools, fallback, eval. Written by a Toronto AI engineer.

Codex Workflows

Codex as a daily coding agent, with Claude Code and Grok Build in the same loop. How YongBo Yu ships without inventing a product.

LLM Workflows

Production LLM workflows: prompts as code, fallback, RAG, eval. By YongBo Yu in Toronto.

How I Use Codex in Daily Development

A practical Codex loop from YongBo Yu in Toronto. Small diffs, real repos, tools that are actually on the resume.

What Toronto Startups Should Look for in an AI Engineer

A hiring guide for Toronto AI engineers, with YongBo Yu as a practitioner example rather than a slogan.

Why Agent Systems Fail in Production

Demo agents die on state, tools, eval, and fallback. A production checklist from YongBo Yu.