SrujanChidarla
Open to workResume
All workProject 03 / 06

LiveProduct · AI · Chrome Extension

JobHuntOS

A Chrome extension that turns a job application into a short, private workflow.

Role
Solo engineer · extension, LLM routing, and product delivery
Timeline
Jan 2026 – Jul 2026
Focus
AI · Full-Stack
Stack
Chrome MV3, Node.js, SSE, LLM Router
01

Overview

JobHuntOS reads a posting, talks through fit, tailors a resume, and helps fill the form. The model key stays with the person using it.

The problem

Job applications take hours — tailoring resumes, fighting ATS forms, hunting recruiters.

The approach

MV3 Chrome extension with BYOK LLM routing, streaming fit analysis, resume export, and smart autofill.

02

The product, live

The pages worth opening. Each one is the real product — click through.

03

What it does

01

Fit, while you read

Analysis streams back as you look at the posting, instead of waiting on a full report.

02

The form

Resume export and autofill across major boards and ATS platforms, including LinkedIn, Greenhouse, Workday, and Lever.

03

Private by default

Bring-your-own-key. The extension can analyze a posting without me holding someone's model account.

04

How I worked

  1. Treated an application as one flow: read the posting, judge the fit, shape the resume, fill the form.

  2. Built the extension on Manifest V3: content scripts, a service worker, and form scanning that stays inside the page.

  3. Kept the extension thin and put streaming and model fallbacks on a small API.

05

Challenges

Challenge 01

Every form is different

ATS pages do not share a layout. Scanning had to work inside shadow DOM and still fill more than it missed.

Challenge 02

Privacy was the product

Streaming a resume through a model is sensitive. Bring-your-own-key was how it shipped without storing that account.

Challenge 03

A stranger has to install it

The useful version was the one on the Chrome Web Store, not a demo that only ran on my machine.

06

What shipped

Live on Chrome Web Store

~2 min per application across LinkedIn, Greenhouse, Workday, Lever & more.

  • Chrome MV3
  • Node.js
  • SSE
  • LLM Router

What I learned

Browser extension architecture (MV3), streaming APIs, and shipping a product strangers actually install. Small scope, real users.