Skip to content
View Jair-pc's full-sized avatar

Block or report Jair-pc

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Jair-pc/README.md

Jair Pereira da Silva Júnior

AI Engineer | Data Engineer | Analytics

Building AI applications in production and scalable data pipelines

LinkedIn Portfolio Email WhatsApp


About Me

AI Engineer and Data Engineer with 4+ years of experience developing AI applications in production and building data pipelines and analytical solutions. Based in Belo Horizonte, MG — Brazil.

Currently pursuing a postgraduate degree in Generative AI & LLMs at PUC Minas.

  • AI Engineering in production: LLMs, autonomous agents, RAG systems, API integrations, observability and continuous improvement
  • Data Engineering: ETL/ELT pipelines, medallion architecture, dbt, Databricks, Snowflake
  • Analytics & BI: Power BI, SQL, Python, data modeling and dashboards
  • Projects delivered for AgroTech, banking (Santander, Itaú Unibanco via Aris Tecnologia) and SaaS clients

Tech Stack

AI Engineering

Python FastAPI Anthropic OpenAI Google Gemini LangChain n8n

Data Engineering

Apache Spark Databricks Snowflake dbt Airbyte Apache Airflow

BI & Visualization

Power BI Streamlit Metabase

Databases & Infrastructure

PostgreSQL MySQL MongoDB Redis Docker Azure


AI Projects in Production

Real applications running with real users — codebases are private (client work).

+Leite AI Platform — Micro SaaS for Milk Quality

Specialized AI platform for milk quality consultants: domain-specific agent with RAG context retrieval, real-time analytical dashboards, automated technical recommendations and structured data analysis.

  • Stack: Python · FastAPI · PostgreSQL · Docker · OpenAI · Anthropic · Gemini · RAG · LangChain · Prompt Engineering
  • Highlights: Production RAG pipeline, multi-LLM integration, observability, real-time dashboards, scalable SaaS architecture

Lumax WhatsApp Agent

Conversational AI agent integrated with WhatsApp: automated customer service with persistent conversation history, context-aware responses via LLMs, and continuous prompt refinement.

  • Stack: Python · FastAPI · Claude API · PostgreSQL · Docker · REST APIs · Prompt Engineering
  • Highlights: Persistent conversation context, LLM-powered responses, latency monitoring, observability, production-grade deployment

Featured Open-Source Projects

⭐ LH Nautical — End-to-End Data Platform

From 4 messy raw files to a complete analytics platform: medallion architecture (Bronze → Silver → Gold), Kimball star schema, sales & customer analytics, demand forecasting, recommender system and a live web app. R$ 2.61B in revenue analyzed.


BanVic 360 — Lakehouse with 9 ETL/ELT Stacks Compared

Complete lakehouse project: SQL, Python, Apache Hop, Docker, Airflow, dbt, Databricks, n8n and Microsoft Fabric — all compared against the same ground truth. 50,997 clients · 3M+ transactions · 8/8 KPIs validated.

  • Stack: SQL · Python · dbt · Databricks · Microsoft Fabric · Power BI
  • 📂 Repository · 📖 Case study

UCBvet — AI Support Chatbot (RAG + SQL)

Generative AI chatbot for a bovine insemination company: RAG over domain documents combined with SQL lookups, powered by Groq LLM.


RAG Chatbot — Chat With Your Documents

Upload PDFs and TXT files, ask questions in natural language: embeddings + vector search with ChromaDB, served in a Streamlit UI.


More Public Projects

Series Repositories
SQL Kickstarter — 370k projects · Dimensional modeling (Kimball) · Sales analysis (CTEs + window functions) · RFM segmentation (NTILE)
Pandas / Analytics Fraud detection — 1.3M transactions · Cohort retention · ABC inventory — 32k SKUs · Delivery delays — 96k orders
Power BI Game Sales · ProUni · Production / OEE · HR / turnover · People Analytics
Machine Learning Employability, depression & market basket (PUC Minas)
Data Engineering Databricks medallion pipeline (PySpark + Delta Lake)

📖 Full portfolio with case studies: jrdados.com.br/projetos.html


GitHub Stats

GitHub Streak

Activity Graph


Education

Year Degree Institution
2026 Postgraduate — Generative AI & LLMs (in progress) PUC Minas
2022 Postgraduate — Data Engineering PUC Minas
2018 Bachelor's — Electrical Engineering UNI BH
2015 Technician — Electrical/Electronics UNA
2010 Bachelor's — Advertising & Marketing Estácio de Sá

Selected Certifications

Year Course Institution
2026 Snowflake & dbt (in progress) Arruda Data Consulting
2025 Data Engineer — 424h (in progress) DSA
2025 n8n Automations — 50h Jornada de Dados
2025 Analytics Engineering — 60h Indicium
2024 AI & Data Analysis with LLMs — 160h Uni1500
2022 Bootcamp AWS · Azure · Data Engineering XP Educação

Open to opportunities in AI Engineering, Data Engineering and Analytics.

LinkedIn

Pinned Loading

  1. banvic-analytics banvic-analytics Public

    BanVic Analytics revisado: pipeline ETL em PostgreSQL (stg_banvic), staging antes do Power BI, integracao com IPCA externo. 4 analises SQL (clientes, credito, IPCA, transacoes).

    Python

  2. banvic-v1-powerbi banvic-v1-powerbi Public

    BanVic V1 - entrega original do Desafio Lighthouse 2024 (Power BI). .pbix + relatorio PDF + mockups Figma + Power Query M + DAX documentados.

  3. databricks_emprega_dados_medallion databricks_emprega_dados_medallion Public

    Pipeline medallion (Bronze/Silver/Gold) no Databricks com PySpark + Delta Lake. Pedidos de delivery (EmpregaDados).

    Python

  4. IA-ChatBot IA-ChatBot Public

    Python

  5. lh_nautical_project lh_nautical_project Public

    Projeto end-to-end de dados: 4 bases brutas → arquitetura medallion (bronze/silver/gold) → análises, previsão, recomendação e web app Streamlit ao vivo. R$ 2,61 bi de receita analisada.

    Jupyter Notebook

  6. pandas_deteccao_fraudes pandas_deteccao_fraudes Public

    Analise de 1,3M transacoes de cartao de credito (Kaggle) para identificar onde a fraude se concentra. Pandas + classes desbalanceadas.

    Python