VINICIUS FRANÇAAI / SOFTWARE ENGINEER
Résumé
Independent portfolioSão Paulo, Brazil

AI engineering · Software · Machine learning

VINICIUS
FRANÇA.

I build the productand the systems behind it.

Backend services, AI workflows and applied machine learning — from the first idea to the thing running in production.

  • Backend services
  • Agent workflows
  • Applied ML
Download résumé (PDF)

Selected work
From system to product

  1. 01 / Product · BetaWinperium ↗Habits, routines and AI agentsProduct & AI engineering
  2. 02 / Product · ProductionSiouve ↗Software for psychologyGo backend · Architecture
  3. 03 / Team achievementAWS + TiDB ↗Builders Day hackathon winnerRetrieval · Risk models
Next ↓The person00 — 05

01The person

Identity · Practice

The person.The practice.

Builder by practice. Athlete by discipline.

I'm Vinicius. I study Software Development at FATEC Itaquera and build across AI and software engineering. I learn by building complete products: the idea, the architecture, the backend and the intelligence inside it.

Before engineering, there was competitive volleyball. Professional beach volleyball taught me to repeat, read the feedback and adjust. I bring that same discipline to the systems I build.

Build. Test. Read the feedback. Adjust.

A shared method
Fig. 02
Fig. 01
Currently studying
Software Development · FATEC Itaquera
Beyond the screen
Professional beach volleyball
On the court
Centro Olímpico de Treinamento e Pesquisa · State and national competitions
Next — 02Selected work ↓

02Selected work

04 case studies

Ideas,in operation.

Three products in use. One classical ML study.

Inside each case: the problem, the architecture and the decisions behind the implementation.

Drag to explore / Open to inspect

01 / 04

Beta / Production

Winperium

Routines, habits and goals, with Alfred, an AI coach that never changes anything silently.

  1. 01

    Winperium

    Context, retrieval, structured agents.

    Product engineering ↗
  2. 02

    Siouve

    Professional workflows. A Go backend in production.

    Software architecture ↗
  3. 03

    SupBetter

    Intent before identity. Go owns the rules.

    Conversational AI platform ↗
  4. 04

    Asteroid risk

    Classification. Cross-validation. Comparison.

    Classical machine learning ↗
Next — 03Tools for the work ↓

03Toolkit

Applied / Ongoing study

Tools forthe work.

The toolkit follows the problem.

Fig. 03

Three ways of deciding

  1. 01

    Rules

    Deterministic logic stays in code, where it can be read and tested.

    Software
  2. 02

    Prediction

    Models are compared and evaluated before they are trusted.

    Machine learning
  3. 03

    Interpretation

    Agents operate with context, tools and explicit validation.

    AI engineering
01

Software engineering

Go / Python / TypeScript / FastAPI

In practice

The services behind the product.

APIs, authentication, migrations and service/repository boundaries. The work includes deployment, integration and production debugging.

  • Go
  • Python
  • TypeScript
  • FastAPI
  • PostgreSQL
  • SQLAlchemy
  • Docker
02

AI engineering

LangGraph / RAG / Embeddings / Vector retrieval

In practice

AI workflows as software systems.

Explicit agent state, intent routing, context management and tool orchestration. Deterministic rules remain in deterministic code.

  • LangGraph
  • RAG
  • Embeddings
  • Vector retrieval
  • Structured outputs
  • Validation
03

Machine learning

scikit-learn / pandas / NumPy / Cross-validation

Projects & study

Prepare. Train. Compare. Evaluate.

Classical classification and model comparison in the asteroid project. Broader studies include regression, clustering and anomaly detection.

  • scikit-learn
  • pandas
  • NumPy
  • Cross-validation
  • Feature engineering
  • Model evaluation
04

Deep learning / CV

PyTorch / OpenCV / Neural networks / Autoencoders

Active study

The next layer of the practice.

An active learning direction. Earth Report is the intended project for deeper ML, Deep Learning and Computer Vision work; that layer is still in development.

  • PyTorch
  • OpenCV
  • Neural networks
  • Autoencoders
  • Computer vision
05

Product frontend

React / Next.js / TypeScript / GSAP

In practice

How the system is experienced.

Responsive product interfaces and interactive experiences. This portfolio is part of that engineering practice.

  • React
  • Next.js
  • TypeScript
  • GSAP
  • CSS
06

Engineering workflow

Linux / Fedora / Git / GitHub / Docker

In practice

Build, inspect and iterate.

Daily development on Linux, version control, local environments and production debugging. AI-assisted tools support the workflow alongside product and architecture decisions.

  • Linux / Fedora
  • Git
  • GitHub
  • Docker
  • VS Code

Education / In progress

FATEC Itaquera

Multiplatform Software Development

Verified certificates

Next — 04Evidence ↓

04Evidence

Award · Production · Study

Evidence,not adjectives.

AWS + TiDB / Builders Day

Hackathonwinner

Part of the team behind a flight operational-risk and financial-exposure system. Retrieval, predictive models and deterministic cost calculations feed an AI analyst.

  • LangGraph
  • TiDB
  • Python
Production

Winperium ↗Siouve ↗

Ongoing studyAI Engineering Senior ↗Structured learning, documented in code.
Global submissionAI Tinkerers ↗SupBetter · Result pending

Inside the winning system / Flight risk & financial exposure

  1. 01

    Retrieve

    Historical patterns and similar cases.

    Vector retrieval
  2. 02

    Calculate

    Operational risk and financial exposure.

    Deterministic code
  3. 03

    Interpret

    An AI risk analyst reading the retrieved context.

    LangGraph
  4. 04

    Validate

    Structured output and explicit validation.

    Schemas
Next — 05Let's talk ↓

05Contact

Open to conversation

LET'S BUILDSOMETHING.

AI, software, products. A conversation is a good place to start.

Write to me
Download résumé (PDF)

Direct channels

  • Product engineering
  • Agent systems
  • Applied ML
Practice
AI engineering · Software · Applied ML
Study
Software Development · FATEC Itaquera
Base
São Paulo, Brazil
Built with
Next.js · React · GSAP