# João Rebeschini

João Rebeschini is a remote worker on Himalayas, Founder & AI/IoT Engineer at TerraLink, based in Brazil.

Source: https://himalayas.app/@joorebeschini

```text
I build Generative AI pipelines for complex document and agricultural intelligence workflows.
```

| Detail | Value |
| --- | --- |
| Availability | Looking for a job |
| Location | Brazil |
| Authorized to work in | Brazil |
| Salary expectations | 30k+ USD |
| Job categories | AI Engineer, Data Scientist, Machine Learning Engineer, IoT Engineer |
| Skills | Generative AI, Large Language Models, Prompt Engineering, Structured Extraction, Document Classification, LLM Evaluation, Human In The Loop, OCR Pipelines, LangGraph, Lang Chain, Jinja, scikit learn, TensorFlow, Keras, YOLO, Detectron2, Python, SQL, Pandas, NumPy, PySpark, PostgreSQL, Git, REST APIs, LLM APIs, Testing Automation, Prompt Versioning, Data Pipelines, GPT, Claude, Gemini, IoT, LoRaWAN, NB IoT, Cloud Infrastructure, Data Lake, Data Warehouse, Apache NiFi, Google Cloud Dataflow, Pentaho, Power BI, Web Scraping, Data Cleaning, Data Standardization, Bag Of Words, Naïve Bayes, Random Forest, SVM, Deep Learning, Computer Vision, Faster RCNN, RetinaNet, Object Detection |
| LinkedIn | https://linkedin.com/in/pedro-rebeschini |

## What I'm looking for

```text
I'm looking to build applied AI systems that combine LLMs, machine learning, data pipelines, and real-world workflows, especially where reliable automation and human-review feedback improve complex decisions.
```

## About

```text
At Elogroup, I build Generative AI pipelines that draft administrative tax litigation reports from OCR-processed legal documents.

I design multi-stage workflows for classification, structured extraction, deterministic processing, report compilation, and LLM-based text refinement. I use Python, LangGraph, LangChain, Jinja templates, versioned prompts, and automated tests to improve accuracy, completeness, and consistency.

I'm also building TerraLink, an intelligent soil-monitoring platform that connects field sensors, LoRaWAN/NB-IoT, cloud processing, and AI to support irrigation and crop-management decisions.

My foundation in data science includes Python, PySpark, SQL, data governance, machine learning experiments, and computer vision for aerial agricultural imagery. I've worked with GPT, Claude, Gemini, YOLO, Faster R-CNN, and RetinaNet across applied enterprise and agricultural use cases.
```

## Experience

### Founder & AI/IoT Engineer (current)

TerraLink. Feb 2026 - Present

```text
Building an intelligent soil monitoring platform combining IoT, real-time data, cloud infrastructure, and Artificial Intelligence. Designed data architecture from field sensors and LoRaWAN/NB-IoT connectivity to cloud processing, storage, and application layers.
```

Skills: IoT, LoRaWAN, NB-IoT, Cloud-Infrastructure, AI, Data-Architecture, Time-Series-Analysis, Alerts, Indicators

### Data Scientist / AI Consultant (current)

Elogroup. Apr 2024 - Present

```text
Develops Generative AI pipelines for automated drafting of administrative tax litigation reports from OCR-processed legal documents. Designs prompts for document classification and structured legal information extraction, and orchestrates graph-based AI workflows with LangGraph/LangChain.
```

Skills: Generative-AI, Large-Language-Models, Prompt-Engineering, Document-Classification, Structured-Extraction, LangGraph, Lang-Chain, Python, Jinja, OCR, GPT, Claude, Gemini, Testing-Automation, Data-Pipelines

### Data Science Intern

[Embraer](https://himalayas.app/companies/embraer). May 2022 - Oct 2022

```text
Supported data centralization and governance initiatives across Data Lake/Data Warehouse environments. Processed SAP data using Python/PySpark and SQL and developed Power BI dashboards.
```

Skills: Data-Lake, Data-Warehouse, Apache-NiFi, Google-Cloud-Dataflow, Pentaho, Python, PySpark, SQL, Power-BI

### Data Science Intern

Ministry of Science, Technology and Innovation. Feb 2021 - Jul 2021

```text
Developed web scraping, data cleaning, and standardization pipelines and supported Machine Learning experiments using Bag of Words, Naive Bayes, Random Forest, and SVM. Co-authored a scientific paper published at EnANPAD.
```

Skills: Web-Scraping, Data-Cleaning, Data-Standardization, Machine-Learning, Bag-Of-Words, Naïve-Bayes, Random-Forest, SVM

## Education

### University of Brasília

Bachelor of Science, Mechatronics Engineering. 2018 - 2023

```text
B.Sc. in Mechatronics Engineering. Final project focused on Deep Learning and Computer Vision for aerial agricultural imagery, applying YOLO, Faster R-CNN, and RetinaNet for object detection and crop-row analysis.
```

## Tech stack

- NumPy
- Pandas
- PySpark
- PostgreSQL
- Python
- TensorFlow
- scikit-learn
- Keras
- Apache NiFi
- Gemini
- Google Cloud Dataflow
- Jinja
- Git
- SQL
- LangChain
- LangGraph
- Power BI

To contact João, message them from this profile with a Himalayas employer account: [Hire on Himalayas](https://himalayas.app/hire).

## People also viewed

- [Felipe da Motta](https://himalayas.app/@felipedamotta): Senior AI Engineer, Motorola Solutions
- [Matheus Rodrigues](https://himalayas.app/@matheusrodrigues1): AI Engineer, Empirical LLC
- [Açucena Rodrigues](https://himalayas.app/@aucenarodrigues): Founding AI Engineer, Coflint OÜ
- [Francinei Gomes](https://himalayas.app/@francineigomes): Senior Python Developer, Thomson Reuters
- [Matheus Cardoso](https://himalayas.app/@matheuscardoso): Data Scientist, Mercado Livre
- [Filipe Côrtes](https://himalayas.app/@filipecortes): Senior Data & AI Consultant, FC Consultoria E Servicos
- [Guilherme Tolentino](https://himalayas.app/@guilhermetolentino): Legal Consultant, Fundação de Desenvolvimento Gerencial
- [Horst User](https://himalayas.app/@horstuser): Artificial Intelligence Architect, Blick Technologies
- [Micaelly Cristine](https://himalayas.app/@micaellycristine): Data Scientist, Prysmo
- [José Henrique User](https://himalayas.app/@joshenriqueuser): Data Engineer, Econodata