Sujit B.

Founders and teams usually come to me when a manual process has grown too large, a scraper keeps breaking, or several tools need to work together without someone watching them all day. I build AI automation, web scraping systems, agents, APIs, and data pipelines. The goal is simple: remove repetitive work and leave you with something your team can understand, run, and maintain. ✓ 100% Job Success ✓ Top Rated ✓ 70 Upwork jobs and 1,282 hours ✓ 5+ years in automation, scraping, APIs, and backend systems Some of my work is public. StealthPlex is my multi-engine Python project for moving between direct HTTP requests and browser automation. I maintain Google News URL Decoder and Facebook Pages Scraper. I also built DockLift for deploying and managing Docker applications. My client work includes multi-source product and supplier search for RBEX and ProcurX. ● AI Automation & Agents → Agents that call APIs, tools, databases, and business systems → Multi-step workflows with structured output and validation → AI-assisted data extraction, classification, matching, and routing → Bots and internal assistants connected to live information → Fallback logic and human approval when a wrong answer matters → Scheduled and event-driven workflows using APIs and webhooks I do not treat an API call with a long prompt as a finished agent. I build the surrounding workflow that gives it context, checks the result, handles failure, and passes the output to the right system. ● Web Scraping & Browser Automation → Dynamic and JavaScript-heavy websites → Protected targets that need sessions, proxies, or browser fallback → Playwright, SeleniumBase, Scrapy, curl_cffi, Scrapling, and Apify → Pagination, authentication, retries, checkpoints, and interrupted-run recovery → Product, directory, supplier, review, news, and structured business data → Delivery to CSV, Excel, JSON, Google Sheets, APIs, or databases → Existing scraper repair, maintenance, and performance debugging I test difficult websites before promising a full extraction. Some targets only need clean HTTP requests. Others need a real browser or a different collection plan. I would rather prove that early than sell you a confident guess. ● APIs & Backend Automation → FastAPI and Python backend services → REST APIs, webhooks, authentication, and third-party integrations → PostgreSQL, Redis, data storage, and background processing → Telegram, Discord, notification, and workflow bots → Data validation, normalization, deduplication, and matching → JavaScript or TypeScript services when they suit the existing stack ● Multi-Source Data Workflows → Collecting results from several websites or APIs → Converting different source formats into one consistent structure → Duplicate detection and source-level error tracking → Scheduled refreshes and change detection → Search, comparison, and supplier or product matching workflows → Clear output that can be reviewed before it reaches production This is the kind of work behind RBEX and ProcurX, where information from different sources has to become one useful workflow rather than a folder full of unrelated files. ● Deployment & Reliability → Docker and Linux deployment → Logs, retries, rate-limit handling, and failure alerts → Environment configuration and secrets handling → Monitoring, restart behavior, and run documentation → Clean handover so the system does not depend on my laptop ● How I Work → Understand the current process and where it fails → Test the riskiest assumption with a representative sample → Agree on the smallest useful milestone → Build the complete workflow after the approach is proven → Keep progress and technical tradeoffs clear → Deliver code, setup instructions, and documentation ● Best Fit → You want to replace repetitive manual work → Your scraper works sometimes but cannot be trusted → You need an AI agent connected to real tools and data → Several APIs, websites, spreadsheets, databases, or bots need to work together → You need the backend and deployment, not only a quick demo → You want honest feasibility advice before committing to a full build Send me the process, website, or API you want to automate and tell me where it currently breaks. I will review the difficult part first and tell you whether it needs a focused fix, a rebuild, or a short paid feasibility test.