You will get AI-Powered Business Automation with n8n, LLMs and API Integrations
Project details
I build AI automation workflows that actually run in production, not demos, not prototypes.
Using n8n for orchestration and LLM agents (Claude, GPT-4o) for the intelligence layer, I design systems that handle email triage, document processing, data extraction, and multi-step business logic without manual intervention.
What sets me apart: 12 years of production software engineering means every workflow I build includes proper error handling, retry logic, logging, and monitoring. My open-source Financial Intelligence Agent, built on LangGraph with multi-model OCR, structured data extraction, and persistent state, is a public example of the architecture I bring to every project.
I'm pragmatic. If your problem is better solved with rules or a simple API call, I'll tell you that. If it genuinely needs an AI agent, I know how to build it properly.
Using n8n for orchestration and LLM agents (Claude, GPT-4o) for the intelligence layer, I design systems that handle email triage, document processing, data extraction, and multi-step business logic without manual intervention.
What sets me apart: 12 years of production software engineering means every workflow I build includes proper error handling, retry logic, logging, and monitoring. My open-source Financial Intelligence Agent, built on LangGraph with multi-model OCR, structured data extraction, and persistent state, is a public example of the architecture I bring to every project.
I'm pragmatic. If your problem is better solved with rules or a simple API call, I'll tell you that. If it genuinely needs an AI agent, I know how to build it properly.
AI Algorithms
Large Language Model, Transformer Model, Variational Autoencoder, YOLOAI Applications
AI Chatbot, AI Content Creation, AI Mobile App Development, Anomaly Detection, Image Analysis, Image Processing, Image Recognition, Object Detection, Sentiment Analysis, Text RecognitionAI Development Language
PythonAI Tools
Azure OpenAI, GitHub Copilot, Hugging Face, PyTorch, Streamlit, TensorFlow, Word2vecAI Models
ChatGPT, DALL-E, GPT-3, GPT-4, LLaMA, Midjourney AI, Naive Bayes Classifier, WhisperWhat's included
| Service Tiers |
Starter
$1,500
|
Standard
$3,500
|
Advanced
$7,000
|
|---|---|---|---|
| Delivery Time | 7 days | 14 days | 30 days |
Number of Revisions | 2 | 3 | 5 |
AI Model Integration | |||
Batch Normalization | - | - | - |
Database Integration | - | ||
Detailed Code Comments | |||
Image Upscaling | - | - | - |
MLOps | - | - | - |
Model Deployment | - | - | - |
Model Documentation | - | ||
Model Monitoring | - | - | - |
Model Testing & Optimization | - | - | |
Model Tuning | - | - | - |
Natural Language Processing | |||
NLP Tokenization | - | - | - |
Pre-Training | - | - | - |
Prompt Engineering | |||
Setup File | |||
Source Code |
Optional add-ons
You can add these on the next page.
Fast Delivery
+$1,000 - $3,500
Additional Revision
+$250
Ongoing Maintenance & Monitoring
(+ 10 Days)
+$500
Additional Tool Integration
(+ 10 Days)
+$300About Leandro
AI/LLM Engineering Specialist
Montevideo, Uruguay - 7:57 pm local time
I design and build LLM-powered agent systems, the kind that go beyond a chatbot wrapper and actually solve operational problems. Multi-step pipelines, tool-calling agents, structured data extraction, human-in-the-loop workflows.
My most recent project is an open source finance intelligence agent built on LangGraph: three independent graphs (reconciliation, insights, chat) sharing a persistent database layer, with tiered duplicate detection, vision-model OCR, LLM accuracy evals, and full CI/CD. It's the kind of architecture I bring to client work, production patterns, not prototypes.
What I work with:
- LangGraph, LangChain, Python, Pydantic, SQLAlchemy
- OpenAI and Anthropic APIs (GPT-4o, Claude Sonnet)
- RAG pipelines, embeddings, vector databases
- Streamlit for rapid UI prototyping
- Computer Vision (co-founded a CV startup, built CNN-based detection systems)
I also bring 12 years of production software engineering (Android/Kotlin), which means I think about error handling, caching, testing, and deployment, not just the LLM call.
I'm pragmatic about AI. If your problem is better solved with rules or a simple API call, I'll tell you that. If it genuinely needs an agent with tool use and structured reasoning, I know how to build it properly.
Based in Uruguay (GMT-3). Good overlap with US and European hours.
Milestone-based delivery. Clear communication about what's feasible and what's risky.
Steps for completing your project
After purchasing the project, send requirements so Leandro can start the project.
Delivery time starts when Leandro receives requirements from you.
Leandro works on your project following the steps below.
Revisions may occur after the delivery date.
Initial MVP
Pepare an MVP with core functionality and send it to the client for testing.
Adjustments to to the MVP
Update the system, make any corrections and iterate again with the client.