CS student at CSULB with an ML specialization, building at the intersection of Agentic AI, Big Data, and real-world product. I'm a 3× hackathon & datathon winner, Co-Founder of an AI startup, and ML researcher extending multi-agent self-improvement architectures. When I'm not shipping code, I'm organizing BeachHacks — CSULB's flagship hackathon and leading the Google SWE Fellows track for the next generation of engineers.
class Phu:
def __init__(self):
self.name = "Quach Thien Phu"
self.role = ["Software Engineer", "AI/ML Engineer", "Startup Founder"]
self.university = "Cal State Long Beach (CSULB) — CS, ML Specialization"
self.grad_year = "May 2027"
self.company = "Co-Founder & COO @ VoiceBridge"
self.internship = "Data Engineer Intern @ OneHope, Inc. (Summer 2026)"
self.leadership = ["ACM VP", "BeachHacks Director", "Google SWE Fellows Lead"]
self.research = "CGEV Multi-Agent Self-Improvement (arXiv:2502.04780 extension)"
self.wins = ["USC AI Challenge Hackathon", "LMU Business Challenge Datathon", "CSULB Beach Hacks"]
self.languages = ["Python", "TypeScript", "JavaScript", "C", "C++", "SQL", "Go"]
self.interests = ["Agentic AI", "LLM Fine-tuning", "Computer Vision", "Big Data"]
def current_focus(self):
return [
"CGEV thesis — extending SiriuS multi-agent pipeline on PubMedQA",
"VoiceBridge — AI voice control & desktop automation startup",
"Data Engineering with Dagster, dbt, and cloud pipelines",
]
def fun_fact(self):
return "I love building projects and competing in hackathons! Gym rat +3 year"Languages
AI / ML & Data Science
Data Viz & Analysis
Databases
Web Frameworks & Backend
DevOps, Cloud & Data Engineering
|
AI Voice Control & Desktop Automation Startup Co-Founder & COO. Won 1st Place @ USC Hackathon (Dain's Challenge). Competing in Sunstone Innovation Challenge.
|
Multi-Agent Self-Improvement Architecture Extending SiriuS (arXiv:2502.04780) with a novel Critic-Guided Evolutionary framework. Fine-tuning on PubMedQA with GPT-4.1-nano via OpenAI API.
|
|
Glitch @ UCLA Hackathon — Mar 2026 Transforms story briefs into cinematic teaser videos using Google GenMedia APIs — Gemini 3.1, Veo 3.1, Lyria RealTime. Submitted to Kaggle competition.
|
BeachHacks — Multi-Agent Study Companion AI-powered study companion using Fetch.ai uAgents framework. Orchestrates multiple specialized agents for adaptive learning.
|
|
UCI Hackathon — Crisis Support Platform Real-time crisis support using Twilio, AssemblyAI, GPT-4o, and ElevenLabs for voice-first mental health assistance.
|
UCSB Hackathon — Best Dining Hall App AI-powered UCSB dining hall app with real-time wait time prediction using YOLO v3 computer vision on live camera feeds, plus personalized menu recommendations based on dietary needs.
|
|
Top 2 — LMU Datathon 2025 Analyzed 25,052 shifts for compliance with California's Wage & Hours law (Section 512a). Identified $232,401 in financial exposure from non-compliance across Company XYZ's workforce data.
|
UCI Datathon ML model analyzing 54,681 Uber driver signups to predict first-trip completion (only 11.22% convert). Identifies high-potential drivers and provides actionable recommendations to improve conversion rates.
|
|
20 Projects — Fundamentals to Production AI A comprehensive collection spanning the full ML stack: neural networks from scratch, computer vision (CNNs, U-Net, FaceNet, MobileNet), NLP & Transformers, Generative AI (LLMs, RAG, vector search), and Agentic AI (CrewAI, LangGraph, Fetch.ai). Each project includes detailed documentation and Jupyter notebooks.
|
|
