PROJECT 03 · Automation · Technical development

Live Stream Automation
From repetitive checks to a connected workflow

At All4e, I built a three-stage automation pipeline that collected live TikTok stream entities, checked profiles, and sent tailored messages to eligible contacts.

ALL4E / THREE-STAGE PIPELINE

COLLECT

VERIFY

NOTIFY

Bhanu Teja Malineni03 / 04

RoleSole developer · Working Student, Technical Developer

PeriodJul—Dec 2025

ContextAll4e ↗ · Berlin, Germany

FocusAutomation

The idea

The team needed to find relevant live stream entities and verify profiles before sending outreach. I was the sole developer of a workflow that connected collection, validation, and messaging, reducing manual cross-checking.

How it works

A Python scraper collected and deduplicated 100 live TikTok stream entities in 3–5 minutes. A Chromium/Selenium checker then verified and classified 100 profiles in 2–4 minutes. A messaging bot notified eligible entities using 10 category-specific templates. A lightweight monitoring dashboard connected the stages and showed workflow output.

01CollectPython scraper gathers and deduplicates 100 live stream entities in 3–5 minutes.

02VerifyChromium and Selenium check and classify 100 profiles in 2–4 minutes.

03NotifyMessaging bot uses 10 category-specific templates for eligible entities.

Reported outcomes

3 stagesCollection, profile verification, and messaging linked in one pipeline.

100 entitiesCollected in 3–5 minutes with automatic deduplication.

2–4 minTo check and classify 100 profiles.

30+ min → near-instantSource-reported reduction in manual cross-checking.

Timing and workflow results are reported in the source resume. “Near-instant” describes the reduced manual cross-checking step, not a published service-level guarantee.

Tools & methods

PythonWeb scrapingChromiumSeleniumAutomation pipelineMonitoring dashboardPrompt engineering

More about the project

I also applied prompt engineering for image and video generation systems. The source material does not specify the platforms or evaluation metrics, so those details are not claimed here.

What I took from it

The main engineering challenge was connecting separate steps into a maintainable workflow and giving the team a clear view of its output. The work combined scraping, browser-based verification, classification, templated messaging, and monitoring.