Applied AI Engineer

Tony Trieu

I build AI agents that plug into the tools enterprise IT teams already use with approval workflows so nothing happens without the right sign-off and keep them running in production.

About

Who I Am

Computer Engineering grad from UC Riverside, Dean's Honor List.

I got into AI engineering because I wanted to build things that people can actually rely on.

AI Systems
AI agents, agentic workflows, RAG, context engineering, model evaluation
Backend
Python, FastAPI, PostgreSQL, pgvector, REST APIs, Ray Serve
Cloud
GCP, Docker, Supabase, Neon, Render
Languages
Python, TypeScript, C++, CUDA C
AI Models
Gemini 2.5 Pro, Gemini 2.5 Flash, Gemini 3.0 Flash

Experience

Where I've Worked

Chinchill.ai

Current

Applied AI Engineer

March 2026 – June 2026

  • · Started as an intern, shipped a production Jira integration on day one, and was promoted to full-time engineer after my first week
  • · Connected the meta-agent to Jira, JSM, ServiceNow, Freshservice, and Confluence so it could perform agentic IT operations
  • · Extended the messaging layer to Slack, Teams, Google Chat, and GitHub so employees get help without leaving the tools they're already in
  • · Built the memory and knowledge layer so agents find the right answer from internal docs instead of guessing, and remember context across conversations
  • · Engineered HITL guardrails so agents never take action without the right person signing off first
  • · Diagnosed and fixed production failures across the deployment layer, authentication, and enterprise connectors

Python  ·  FastAPI  ·  Pydantic AI  ·  PostgreSQL  ·  Ray Serve  ·  Gemini  ·  GCP

Projects

What I've Built

March 2025 – Present

ScheduleBud

Lead Applied AI Engineer

Live App ↗

An AI scheduling app that automates academic scheduling for college students

  • · Students tell it what they need in plain English and it builds a schedule around their needs.
  • · Every answer comes from real course data. The AI doesn't get to guess
  • · Each student's data is isolated at the database level, independent of anything happening in the application code
  • · Simple questions route to cheaper models, while complex ones route to more capable ones.

TypeScript  ·  Node.js  ·  Supabase  ·  Gemini  ·  PostgreSQL

Spring 2025

QKV Attention Acceleration

CS 147 Final Project

Code ↗

A final project about GPU kernel optimization for the attention operation at the core of every large language model

  • · Dug into how memory access, parallelism, and compute bottlenecks interact in GPU kernels, the same tradeoffs that show up in production AI inference at scale

CUDA C  ·  GPU Architecture

Contact

Let's Talk

I'm open to full-time Applied AI Engineering roles. If you're building something with AI agents, enterprise integrations, or AI infrastructure, I wanna learn more about it.

tonytrieu.dev@gmail.com