Project Description for Developer – Amazon Job Auto‑Picker System Below is the full technical explanation of what I need built. Please read carefully because accuracy and performance are extremely important for this project. 1. Current System (What I Already Built) I created a Chrome extension that automatically picks Amazon jobs. The extension uses: 2 API calls 1 WebSocket connection Static jobId + scheduleId Create Application → WebSocket (pending) → Update Application loop The problem is: Amazon recently added strict rate limits The updateApplication call returns 429 (Too Many Requests) 429 is tracked per candidateId + token + IP as per my prespective Even switching IP + candidateId still causes: High latency More 429 403 503 Auto‑cancelled calls This makes job‑picking slow and unreliable. 2. Goal of the New System I want a high‑speed, low‑latency, zero‑error job picker that can: Avoid 429 completely Keep latency between 50–100 ms Run 100+ accounts in parallel Use different IP + token per call Auto‑login accounts every 2 hours Run 10–40 schedule combinations at the same time Send Telegram notification when a job is picked Deploy on AWS with the fastest possible setup (ECS preferred) Bypass the normal flow and jump directly to the interview page (like other pickers do) This system must run 24/7 without stopping. 3. Required Architecture (My Idea + You can improve it) A. Proxy + IP Rotation We need a high‑quality proxy cluster, not a single proxy. Requirements: Proxy servers located closest to Amazon’s region (Canada) Each API call uses a different IP CreateApplication + WebSocket + UpdateApplication must use same IP Zero 403 errors Zero 429 errors B. Multi‑Account Engine System logs in 100 accounts continuously Tokens expire every 2 hours → system must: Detect expiration Open login window Auto‑login Refresh token Continue job picking No manual work. C. High‑Speed Job Picker Logic For each job: Use unique candidateId + token + IP Run 40 scheduleId combinations in parallel UpdateApplication spam must be 0 ms delay (not even 1 ms) If 429 appears → instantly switch IP + candidateId Must maintain 50–100 ms latency D. Direct Interview Page Bypass Other pickers skip the normal flow and go straight to the interview page. I need the same logic implemented. E. JobId + ScheduleId Logic JobId is fixed per station (I already have the full list for Canada) ScheduleId increases weekly in a series Example: This week: 4913 → 5000 Next week: 5001 → 5010 I already have code to generate scheduleId ranges and export to Excel You will integrate this logic into the system. F. Telegram Notification When a shift is picked: Send message to Telegram Include: Candidate name Job name JobId ScheduleId Timestamp G. AWS Deployment Deploy the entire system on AWS: ECS (fastest) Auto‑scaling Low latency networking Proxy integration 4. What I Expect From You You must: Build a fully working, stable, accurate system Test it on the USA Amazon hiring website (many jobs available) Ensure no 403 Ensure no 429 Ensure latency stays under 100 ms Ensure direct interview bypass works Ensure 100 accounts run continuously Ensure 40 schedule combinations run in parallel Ensure each call uses different IP + token Ensure CreateApplication + WebSocket + UpdateApplication use same IP