Hi, I'm

Can Dai.

Software Engineer

I build production Generative AI and full-stack web applications at Hughes Network Systems, from RAG-based chatbots and agentic voice AI to mission-critical Django and PostgreSQL platforms on Azure.

View resume LinkedIn GitHub Get in touch

01 About

I'm a software developer at Hughes Network Systems, where I ship Generative AI features and full-stack web applications that thousands of people rely on every week.

I hold an MS in Computer Science from DePaul University, with research experience in robotics, distributed systems and machine learning. Before that, I earned a BS in Mechanical Engineering from Purdue University, which is where I learned to think in systems and to communicate across disciplines.

Tools I use daily

Python
Django
PostgreSQL
JavaScript
HTMX
Azure
RAG / LLMs
n8n
REST APIs

02 What I do

Generative AI & RAG

Co-developed AVA, a production RAG-based AI chatbot handling 7,400+ weekly sessions with a 40% containment rate, saving an estimated $500K a year.

Full-stack web

Sole engineer behind ATLAS, a mission-critical Django/PostgreSQL app on Azure for 200+ call-center supervisors, with JWT-secured REST APIs and real-time outage monitoring.

Research

Published a paper on robotic swarm contracts at the 2022 IEEE International Conference on Blockchain.

IVR & automation

Modernized IVR branches with n8n pipelines that automate workflows and reduce live-agent call volume.

Chrome extensions

Built RapidRead, an end-to-end Chrome extension that helps you read any text up to 3× faster.

Education

MS Computer Science, DePaul University (2023).
BS Mechanical Engineering, Purdue University (2020).

03 Projects

Personal projects

Featured · Agentic voice AI SaaS · 2025–2026

Customer Service AI Agent (CSAA)

A multi-tenant platform for building, deploying and monitoring AI agents that answer real phone calls, WhatsApp threads and SMS on behalf of a business, with a live dashboard where a human can watch any conversation, take it over, and hand it back to the agent mid-call.

Async Django (ASGI) backend that coordinates Twilio telephony, OpenAI reasoning and ElevenLabs voice on every turn of a conversation
Real-time Next.js dashboard over JWT-authenticated WebSockets, fanned out per tenant through Redis
Celery workers for analytics roll-ups, plus usage metering for every token, voice character and SMS
Graceful fallbacks at every external API, so a vendor outage never drops a live call