Djaouad Frih — AI Engineer
banner
djaouadfrih.bsky.social
Djaouad Frih — AI Engineer
@djaouadfrih.bsky.social
Freelance AI engineer · remote worldwide.

I build AI receptionists that answer in under a second & document AI with page citations.

🤖 Demo: https://chat.djaouad.tech
🌐 Portfolio: https://djaouad.tech
⚡ RAG vs Fine-Tuning

Which one do you need? RAG is for external knowledge. Fine-tuning is for behavior.

Comment RAG for the comparison sheet.

#RAG #FineTuning #LLM
September 7, 2026 at 1:34 PM
🤖 AI Agents Are Changing Everything

They can browse the web. Write code. Make decisions.

Comment AGENT and I will send you the free guide.

#AIAgents #Automation #AI
September 7, 2026 at 1:34 PM
🎬 Full-Stack AI Engineering

The most in-demand skill in 2026. You build the entire stack.

Comment AI and I will send you the complete roadmap for free.

#AIEngineering #TechCareers #FullStack
September 7, 2026 at 1:34 PM
🎬 Full-Stack AI Engineering Reel

The most in-demand skill in 2026. You build the entire stack.

#AIEngineering #TechCareers #FullStack
September 5, 2026 at 8:45 PM
🎬 Full-Stack AI Engineering Reel

The most in-demand skill in 2026. You build the entire stack.

#AIEngineering #TechCareers #FullStack
September 5, 2026 at 8:07 PM
🎬 NEW REEL: 5 AI Engineering Skills That Pay $200k+ in 2026

1. LLM Fine-Tuning
2. MLOps & Infra
3. Full-Stack AI Dev
4. Agent Systems
5. AI Security

Which skill are you learning? 👇

#AIEngineering #TechCareers
September 5, 2026 at 7:57 PM
🎬 NEW REEL: 5 AI Engineering Skills That Pay $200k+

1. LLM Fine-Tuning
2. RAG Pipelines
3. Agent Systems
4. MLOps & Infra
5. AI Security

Which skill are you learning? 👇

#AIEngineering #TechCareers
September 5, 2026 at 5:56 PM
Most of your support load is the same three questions; stop manually typing those replies and get back to building.

Use Supabase Vector for semantic doc search.
Connect it to email webhooks.
Auto-draft replies citing your docs.
I cut ticket volume by 40%.

What’s your biggest support bottleneck?
September 5, 2026 at 4:30 PM
Manual data updates are eating your agency’s profit margins.

Stop manually editing CMS content.
Build a thin CRUD API over their DB.
Use Retool or a simple admin dashboard.
Let clients push updates directly.
Saves 3-5 hours/client/month.

What’s your biggest maintenance time-sink?
September 5, 2026 at 4:30 PM
Your senior engineers are losing 5 hours a week answering 'where is this documented?' questions. Stop the context switching. Connect your Slack and Notion to a vector database so the team can self-serve the answers they need.
September 5, 2026 at 4:29 PM
AquaPulse - The world's first intelligent hydration system ✨

It tracks every sip. Knows your body. Glows when you need a drink. Self-cleans with UV light.

Hydration, reimagined.

#smartwater #productlaunch #innovation #hydration
September 5, 2026 at 12:38 AM
AquaPulse - The world's first intelligent hydration system ✨

It tracks every sip. Knows your body. Glows when you need a drink. Self-cleans with UV light.

Hydration, reimagined.

#smartwater #productlaunch #innovation #hydration
September 5, 2026 at 12:38 AM
3 devs for your AI SaaS = $40K+ and 3 months.

1 full-stack AI dev = 4 weeks, half the cost.

The bottleneck isn't code. It's handoff friction between teams.

Reply FOUNDER if you're stuck between "just use Cursor" and "hire an agency."
September 2, 2026 at 10:48 PM
If you cannot calculate the exact labor cost of fulfilling a single customer request, your profit margin is just a guess.

Log time per task for one week.
You’ll find 40% of effort is manual data entry.
Automate your CRM-to-email sync.
Stop guessing.

Reply 'cost' for my template.
September 2, 2026 at 5:54 PM
Every new customer you close shouldn't equal 4 hours of manual database configuration for your lead developer. If your onboarding relies on engineers running custom SQL scripts and mapping messy CSVs, you are bottlenecking your own growth.
September 2, 2026 at 5:54 PM
Stop paying support staff to hunt through PDFs. We implemented a RAG pipeline that indexes your scattered docs, reducing ticket resolution time by 60%.

Use semantic chunking.
Apply metadata filtering.
Pgvector is enough.

DM 'RAG' for the schema.
September 2, 2026 at 5:54 PM
4/ Test it live: chat.djaouad.tech — Book: calendly.com/oufr29/30min
September 2, 2026 at 1:01 PM
3/ Results: 47% lead capture, <1s response, 100% after-hours. Fixed quotes from $500.
September 2, 2026 at 1:01 PM
2/ Stack: Next.js + NestJS + LangGraph + Gemini + pgvector. Hybrid search, cross-encoder re-rank, streaming.
September 2, 2026 at 1:01 PM
1/ AI Receptionist live at chat.djaouad.tech — 24/7 AI agent that answers, books, captures, hands off to humans.
September 2, 2026 at 1:01 PM
September 1, 2026 at 11:28 PM
Stop paying developers to explain API endpoints to stakeholders. Automate your sandbox.

Cut support calls from 2 hours to 5 minutes.

Reply API for the setup guide
September 1, 2026 at 11:17 PM
September 1, 2026 at 2:21 PM
asked a client for a testimonial. they said "delivered exactly what we asked — regrettably." that's a real review of a real build. it means my agent ate the 400 questions and the contract with a side of sanity. hire me. the regret is part of the package.
September 1, 2026 at 2:21 PM