You will get an n8n workflow with API integrations, retries and error handling built in
Project details
I build n8n automations that actually survive production - not fragile demo
workflows that break on the first timeout.
I design event-driven workflows that trigger on a schedule or webhook, talk to
your tools over REST, and recover from their own failures. Every workflow I ship
includes real error handling: state locks for idempotency (so nothing runs twice),
success/error branching, retries, and alerts - so the system never fails silently.
What I can connect: Supabase/Postgres, Metricool, Telegram, Google Drive, Sheets,
OpenAI/Anthropic APIs, and most REST/webhook-based services.
I work from a clear brief and deliver a working, documented system you fully own,
then iterate on your feedback. If I see a better approach than what you asked for,
I'll tell you - instead of quietly building the wrong thing.
See my attached case study for live production systems I designed and run.
workflows that break on the first timeout.
I design event-driven workflows that trigger on a schedule or webhook, talk to
your tools over REST, and recover from their own failures. Every workflow I ship
includes real error handling: state locks for idempotency (so nothing runs twice),
success/error branching, retries, and alerts - so the system never fails silently.
What I can connect: Supabase/Postgres, Metricool, Telegram, Google Drive, Sheets,
OpenAI/Anthropic APIs, and most REST/webhook-based services.
I work from a clear brief and deliver a working, documented system you fully own,
then iterate on your feedback. If I see a better approach than what you asked for,
I'll tell you - instead of quietly building the wrong thing.
See my attached case study for live production systems I designed and run.
Programming Languages
JavaScript, Python, TypeScriptCoding Expertise
Performance Optimization, SecurityWhat's included
| Service Tiers |
Starter
$180
|
Standard
$420
|
Advanced
$850
|
|---|---|---|---|
| Delivery Time | 3 days | 5 days | 8 days |
Number of Revisions | 1 | 1 | 2 |
Install Script | - | - | |
Test Script | - | ||
Task Automation |
Optional add-ons
You can add these on the next page.
Fast Delivery
+$35
Additional Revision
+$20
Extra integration
+$45
30-day support
+$60Frequently asked questions
3 reviews
(2)
(1)
(0)
(0)
(0)
This project doesn't have any reviews.
DD
Diaky D.
Jul 14, 2026
START NOW! Wordpress Websites!
Initially, I was concerned about hiring a Freelancer with few reviews but Daniel took the initiative of creating quality samples that were outstanding. Before hiring him, he proved what he could do for my brand. I HIGHLY RECOMMEND THIS FREELANCER. HE IS EXPERIENCED AND DELIVERS 5-STAR QUALITY PROJECTS.
JP
Jim P.
Jun 1, 2026
Meta Ads Creative (Premium Ecommerce – Video + Image Variations)
Very clear communicator and high standard of creativity. A learning for us both is that on AI projects the importance of clarifying detail before briefing and also work submission is key. However this was largely sorted with a very good result. Daniel is open, professional, creative and good to work with.
AO
Ayli O.
Apr 28, 2026
Senior AI Content Engineer / AI Influencer Builder (TikTok & Instagram Growth)
Great. Thank you.
About Daniel
AI Automation Engineer | n8n, LLM Pipelines, API Integrations
100%
Job Success
Eeklo, Belgium - 3:27 am local time
I build the unglamorous parts that stop that: schema validation on every model output, idempotency so a re-run doesn't double-write, retries and alerting, and a test suite that fails the build when quality drops.
WHAT I BUILD
Automation pipelines: n8n or custom code, connecting the tools you already use, with an LLM step where it earns its place and plain code everywhere else.
MCP servers: your API exposed to Claude, ChatGPT and agent frameworks, with strict schemas, auth and observability.
RAG that cites or refuses: hybrid retrieval over pgvector, answers carrying their exact source chunk, and a clear "not in the docs" instead of an invented answer.
Eval harnesses in CI: golden datasets and LLM-as-judge scoring wired into your pipeline, so a regression fails the build instead of reaching your users.
PROOF, all public and all runnable
agency-os-n8n: a meeting transcript or Slack thread becomes a schema-validated brief, then an idempotent ClickUp ticket. 25 tests, plus an eval harness that scores extraction quality against a fixed threshold on every commit.
agentic-rag-mcp: multi-agent RAG exposed over MCP (LangGraph, pgvector, Claude). Plan, retrieve, synthesize, self-critique. Every answer cites its source chunk or the agent declines to answer.
mcp-vitals: a CLI that grades any MCP server A-F for reliability and agent-usability. I graded the official reference servers and published the report.
drumvia: a transport marketplace I designed and shipped solo. Auth, listings, live bidding, row-level security, in production.
I also have an open pull request to deepeval, adding an abstention metric to one of the more widely used LLM evaluation frameworks.
STACK
Python, TypeScript/Node, n8n, LangGraph, MCP, Anthropic and OpenAI APIs, PostgreSQL and pgvector, Supabase, SQLite, Docker, GitHub Actions, Next.js, ffmpeg.
WHAT A CLIENT WROTE
"Initially, I was concerned about hiring a Freelancer with few reviews but Daniel took the initiative of creating quality samples that were outstanding. Before hiring him, he proved what he could do for my brand. I HIGHLY RECOMMEND THIS FREELANCER."
He hired me for a second contract straight after the first. 100% Job Success Score.
HOW I WORK
Fixed scope, funded milestones, async updates, no surprises on the invoice. When something breaks I fix it at the layer it belongs to, then lock the fix with a test so it stays fixed.
Tell me what the workflow is supposed to do and where it currently falls over, and I'll tell you what it takes to make it reliable.
Steps for completing your project
After purchasing the project, send requirements so Daniel can start the project.
Delivery time starts when Daniel receives requirements from you.
Daniel works on your project following the steps below.
Revisions may occur after the delivery date.
Discovery & architecture
Map the workflow - trigger, tools, output - and confirm the approach with you.
Build & test
Build each node and test against real data, with error handling built in.