ZainJaafil


I build software that other people's businesses actually run on.

Voice agents, internal dashboards, automation. Two of the systems I built handle live sales calls for a payments company right now, on a server I set up myself.

Before any of that I spent a summer cold calling auto shops and nail salons across Orange County trying to sell them websites. Some of it worked.

School
Legacy Magnet Academy, 12thDual enrolled at IVC since 9th
Currently
AI product intern, VitalisBusiness ops intern, local gym
Off the clock
UPSL semi-pro soccerHead coach, U6 to U8
I

A working history

Drag to scrub
Fig. 1 Everything I started, and what happened to it. JUN 2023 0 running 0 dead
Running Dead Moment
Click any row for the story
II

Three of them in depth

Longer form
Vitalis Corp
2026 to present
aisales.shift4tech.com
Live

An AI rep that calls small businesses and hands them to a human

Alex calls a restaurant owner about their payment terminal, works out whether they are worth a rep's time, and then does the part most voice bots skip: it asks permission, puts the caller on hold with music, dials a real salesperson, whispers a summary of the conversation into their ear, and bridges the two of them together.

It replaced a stack built on Vapi, which took the per minute platform fee to zero. It runs on the client's own EC2 box under systemd rather than a third party voice platform, so nothing about it depends on a vendor staying cheap.

The model itself remembers nothing between calls. Every session starts blank. What looks like memory is a one paragraph summary the model writes about itself at the end of each call, stored in SQLite and fed back in at the start of the next one. That was deliberate. Feeding whole transcripts back would bloat every call, and keeping it in my own database means resetting someone's history is one line of SQL instead of a support ticket.

  • Twilio
  • OpenAI Realtime
  • Pipecat
  • FastAPI
  • SQLite
  • nginx
  • EC2
Vitalis Corp
2026 to present
Windows desktop
Live

A tray app that listens to sales calls and feeds the rep their next line

Call Assist sits in a rep's system tray and captures two audio streams at once: their microphone, and the customer's voice coming back through system audio. Both go out for live transcription. The moment it hears the word contract, or too expensive, it pops the matching rebuttal card on screen. It tracks what the customer complains about across the call and assembles a closing script as it goes.

The best thing I did on it was notice a mistake I had already made. The transcription API key was hardcoded in the app, which meant it was going to ship inside every copy installed on every rep's laptop, extractable by anyone who looked. I moved the entire audio connection server side so the key never leaves the server, and put auth on both endpoints so nobody could run up the client's bill.

  • Electron
  • Node
  • WebSockets
  • Deepgram
  • Web Audio
  • pm2
Local gym
2026 to present
On the floor
Live

A dashboard nobody asked me to build

I was hired to work the front desk, call leads off Meta ads, and help with marketing. Two hundred calls in, I noticed the gym had TVs on the wall showing nothing. So I built a live floor display: check ins, streaks, goal progress, the class schedule, member shoutouts. Then I pitched it to the owner.

He is buying more TVs to put it on.

III

Rooms I talked my way into

2026

First place, Berkeley Haas Student Association Pitch Competition

Led a team of five to first out of thirty teams in front of a three judge panel. Everyone else competing was in college. I owned the product concept, the deck, and the pitch itself.

2026

Round one winner, Mott Million Dollar Challenge

One of 1,500 selected out of more than 3,700 submissions nationally. I co created the concept and presented on camera: AI glasses that help blind users get around.

2025

OC Maker Showcase

Team of three, concept through final pitch. I did the SolidWorks design work and delivered to the judges.