# matheus sampaio

matheus sampaio is a remote worker on Himalayas, Data AI Engineer at Maggu, based in Brazil.

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

```text
AI/ML Engineer at Maggu who raised medicine recommendation accuracy to 94% with LLM evaluation.
```

| Detail | Value |
| --- | --- |
| Availability | Looking for a job |
| Location | Brazil |
| Authorized to work in | Brazil |
| Salary expectations | 100k+ USD |
| Job categories | AI ML Engineer, Data Engineer, Data Scientist |
| Skills | LLM Evaluation, LLM As A Judge, Tool Calling Evaluation, RAG, MLOps, MLFlow, Databricks, PySpark, Delta Lake, Terraform, Airflow, CI CD For Models, Model Registry, Drift Monitoring, Observability |
| Website | https://mldump.com |
| LinkedIn | https://linkedin.com/in/mnsgr |
| GitHub | https://github.com/mnsgrosa |

## About

```text
At Maggu, I reworked the LLM judge for medicine recommendations, raising accuracy from 62% to 94% with adversarial agents and a human-made evaluation dataset. I also tracked operational costs, latency, accuracy, and routing in Databricks MLflow.

At Maggu, I cut time to first token from 1.5 seconds to 0.7 seconds by revising cache and RAG strategies. I provisioned reproducible Databricks pipelines with Terraform and reduced compute spend by profiling PySpark jobs.

At Capgemini, I wrote an evaluation suite for agent workflows that caught four regressions before they reached the client. I also ran workshops on agent design and LLM-as-a-judge for a newly formed team and two managers from Vivo.

Across my work at Acaso and Di2win, I built RAG and model-training pipelines, automated document ingestion, and shipped generative AI proofs of concept. I bring a computer vision research background from UFPE and have worked across data pipelines, backend systems, and production ML infrastructure.
```

## Experience

### Data AI Engineer

Maggu. Apr 2026 - Aug 2026

```text
Improved medicine recommendation accuracy from 62% to 94% by reworking the LLM judge into two adversarial agents backed by a human-made dataset on MLflow hosted at Databricks. Tracked LLM operational costs, latency, and accuracy, and reworked cache and RAG strategies to cut time-to-first-token from 1.5 to 0.7 seconds.
```

Skills: Databricks, PySpark, Delta-Lake, MLFlow, Terraform, Agno, GitHub-Actions, LLM-Evaluation, RAG, Cost-Attribution

### Data AI Engineer

[Capgemini](https://himalayas.app/companies/capgemini-technology-services). Jan 2026 - Apr 2026

```text
Wrote evaluation suites for agent workflows, catching 4 regressions before client delivery. Ran workshops on agent design and LLM-as-a-judge for a newly formed team and two managers from Vivo, and handled client technical communication in an Agile delivery team.
```

Skills: Python, LLM-Evaluation, LLM-AI-Agent-Frameworks, Tool-Call-Validation, RAG, Agile-Scrum

### Data AI Engineer

Triggo.ai. Dec 2025 - Jan 2026

```text
Engineered prompts to keep LLM tone consistent using self-consistency and tree-of-thoughts techniques. Tuned Anthropic and OpenAI API calls to enforce JSON-formatted, length-constrained outputs.
```

Skills: Python, Bruno, DocumentDB, Prompt-Engineering, LLM-API-Tuning

### Data Engineer

Oncase. Aug 2025 - Nov 2025

```text
Maintained a Flask API for image processing, storing uploads on S3 for a computer vision inference pipeline. Maintained Airflow DAGs using a Gitflow branch pattern and hosted Airflow on EC2 to run continuous inference and training jobs.
```

Skills: Python, Flask, Paddle-OCR, Airflow, AWS, Terraform

### Machine Learning Engineer

Acaso. Jan 2025 - Jun 2025

```text
Automated document ingestion in Airflow, giving the team back roughly 20% of its week. Built a RAG pipeline using Docling for PDF parsing, LangChain for orchestration, and PostgreSQL with pgvector for storage, and cut manual data prep by 60% by moving S3 ingestion into Python and Boto3.
```

Skills: Python, Lang-Chain, PostgreSQL, Pgvector, Docling, Airflow, S3, AWS-SDK-(Boto3), fastAPI, Streamlit

### Machine Learning Engineer

Di2win. Apr 2024 - Jan 2025

```text
Cut energy consumption 20% with LSTM and regularized regression forecasts tuned in Optuna. Replaced manual retraining with event-driven Airflow pipelines and set up GitHub Actions with Pytest coverage on model behavior and input data.
```

Skills: Python, PyTorch, Optuna, Airflow, Docker, GitHub-Actions, Pytest, fastAPI

### Data Scientist

Valorian. Apr 2023 - Apr 2024

```text
Reduced hosting costs 25% by containerizing workloads and moving to reserved instances. Automated ingestion and transformation end to end, saving 30 hours of manual work per month, and improved wheat flour production efficiency 12% with XGBoost models tuned in Optuna.
```

Skills: Python, XGBoost, Optuna, Docker, AWS-EC2-S3, Plotly-Dash

## Education

### Federal University of Pernambuco (UFPE)

Bachelor of Science, Computer Science. 2022 - Present

```text
Pursuing a Bachelor of Science in Computer Science at the Center of Informatics, Federal University of Pernambuco, with expected completion in July 2026.
```

## Tech stack

- Dash
- AWS IAM
- Amazon EC2
- Amazon CloudWatch
- Amazon S3
- Google Cloud Storage
- Kubernetes
- GitHub Actions
- Pandas
- PySpark
- PostgreSQL
- Databricks
- Terraform
- Python
- PyTorch
- MLflow
- scikit-learn
- Streamlit
- Flask
- FastAPI
- AWS Lambda
- pytest
- Docker
- Airflow
- SQL
- Google Cloud Run
- Amazon SageMaker
- Prefect
- XGBoost
- SciPy
- Amazon ECR
- LangChain
- ChromaDB
- Pydantic
- Delta Lake
- OpenAI API
- Bash
- pgvector
- Optuna
- Bruno
- Docling
- Apache Kafka
- AWS

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

## People also viewed

- [Fernando Bascevicius](https://himalayas.app/@fbarscevicius): Tech Lead Machine Learning Engineer, Tricura Insurance Group
- [Ían Muliterno](https://himalayas.app/@anmuliterno): Machine Learning Engineer, Independent Contractor
- [Beatriz Maia](https://himalayas.app/@beatrizmaia): Machine Learning Engineer, Cadastra
- [Adriel Martins](https://himalayas.app/@adrielmartins): Senior AI Engineer, Lean Solutions Group
- [Francinei Gomes](https://himalayas.app/@francineigomes): Senior Python Developer, Thomson Reuters
- [Gustavo Villar](https://himalayas.app/@gustavovillar): Data Scientist Technical Lead, Hubify
- [Leonardo Ribeiro](https://himalayas.app/@leonardoribeiro): AI & MLOps Owner, Natual AI
- [Renan Pinheiro](https://himalayas.app/@renanpinheiro): Senior Machine Learning Engineer, Blend360
- [Luiz Ferreira](https://himalayas.app/@luizferreira1): Digital Solutions Consultant, Capgemini
- [Luiz Gustavo Ferreira Nazareth](https://himalayas.app/@luizgustavoferreiran): Data Engineer, Capco
