An SEO outreach team scraped sites by hand, wrote personalized emails, and lost leads when no one sent follow-ups. At thousands of contacts, the team couldn't keep up.
I built an outreach pipeline: site scraping with domain extraction, personalized email sending via Gmail OAuth, and independent flows that track opens and trigger scheduled follow-ups. I also built a custom CRM with lead tagging to track each lead's status across the funnel.
The pipeline saves 2–3 minutes per contact, runs follow-ups on schedule, and gives the team a single view of every lead's status.
Stack: n8n · Web scraping · Gmail OAuth · Custom CRM · Automated follow-up flows
0
9
A B2B consulting firm handling 20–30 client chats daily had a support bottleneck: repetitive questions, no 24/7 coverage, and leads slipping through the cracks due to slow response times.
I built a fully automated AI Assistant with RAG integration — trained on their internal knowledge base, capable of handling customer inquiries around the clock without human involvement. The system automatically captures qualified leads, routes edge cases to human agents, and sends full chat transcripts with analytics to the team after every session.
Result: instant FAQ resolution 24/7, zero missed leads, and ≈2.3 hours/day saved on manual support.
Stack: n8n · RAG pipeline · OpenAI · Webhook automation · CRM integration
0
15
I built a video localization pipeline that takes a source video and outputs dubbed, lip-synced, subtitled versions across multiple languages, with the pipeline handling everything from ingest to delivery.
A webhook fires when a new video lands in Google Drive. FFmpeg strips the audio. ElevenLabs dubs it per locale. The pipeline time-stretches the dubbed audio to match the original duration, muxes it back in, and hands it to Sync.so (http://Sync.so) for lip resync. GhostCut scrub the original burned-in subtitles. Fal.ai burns new ones in the target language. GPT-4o translates the title; JSON2Video renders the final cut with the overlay. The output lands in a per-locale folder in Drive and fires a Telegram alert.
~2 cost / 1 min video ⚡
0
48
I built an image localization pipeline for marketing teams shipping creatives across multiple markets.
A webhook triggers on new files in Google Drive. An AI model rewrites all in-image text per locale, preserving the original fonts, colors, and layout. The pipeline uploads localized versions back to Drive.
It runs 24/7, processes 50+ images per batch, and handles as many target languages as you configure in a single execution. A localization cycle that took a designer 2+ days per locale now takes 10 minutes.
Last results from client:
300 creatives localized in less than 50 min ⚡
0
64