Portfolio/Auto-Bid
INTERNAL2024

AUTOMATION · INTERNAL TOOL

Auto-Bid

Automated pipeline for winning projects on domestic and international outsourcing platforms, from project collection to analysis and proposals

Auto-Bid

Auto-Bid — 20s Overview

Project Video

Overview

Automatically collects and normalizes new projects from Wishket, Freemoa, and Upwork. Scores based on niche keywords, budget, and complexity, then auto-classifies as BID or REVIEW. Generates proposal drafts, KPI reports, and project logs to dramatically reduce repetitive bidding work.

Our Role

Pipeline Architecture · Crawler/Connector Development · Scoring Algorithm · Admin Dashboard · Automation Scheduling

Tech Stack

PythonPlaywrightAutomationCrawlerKPI Dashboard

The Challenge

  • 1Discovery overhead — manually monitoring 3+ platforms daily
  • 2Evaluation inconsistency — human mood/fatigue affects project selection
  • 3Proposal fatigue — time-consuming, generic templates get ignored
  • 4Anti-bot defenses — IP blocking, Cloudflare, session management

Our Solution

  • 5-stage pipeline: Collect → Normalize → Score → Decide → Propose
  • ML-inspired 4-axis scoring across 84 canonical technologies
  • AI proposal generation — Claude generates platform/language-specific drafts
  • Human-in-the-loop — AI proposes, human approves, Playwright submits

Key Features

5-Stage Pipeline

Collect → Normalize → Score → Decide → Propose

3-Platform Collection

Wishket, Freemoa, Upwork with anti-bot bypass

4-Axis Scoring

Tech match 40% + complexity 25% + client quality 20% + budget 15%

AI Proposal Generation

Claude-powered, platform and language adapted

Architecture

Auto-Bid Architecture

Live Metrics

90+

Daily Projects Monitored

~21%

Intelligent Bid Rate

2-4s

Per Proposal Generation

90%+

Deduplication Accuracy

Want to build something like these?

Let's discuss your project. We'll share relevant experience and propose the right approach.

Discuss Your Project
Auto-Bid – Case Study | DCC Lab