Disponível para Projetos Freelance Available for Freelance Projects

Olá, eu sou Hi, I'm Walter José Horning Junior

Senior |

Transformo dados complexos em impacto real para o negócio. 15+ anos em TI, 8+ em Data Science e IA. De otimizações de rota que reduziram custos em 13% a pipelines 6.000x mais rápidos. Entrego resultados que geram valor. Turning complex data into measurable business impact. 15+ years in IT, 8+ years in Data Science & AI. From route optimization saving 13% in costs to pipelines running 6,000x faster. I deliver results that move the needle.

Ver Projetos View My Work Fale Comigo Hire Me
Walter José Horning Junior
Role Scroll
Sobre Mim About Me

Unindo Data Science
e Impacto no Negócio Bridging Data Science
& Business Impact

0+ Anos em TI Years in IT
0+ Anos em Data/IA Years in Data/AI
0+ Projetos Entregues Projects Delivered
0 Setores Atendidos Industries Served

Sou Senior Data Scientist e AI/ML Engineer, trabalhando 100% remoto (de qualquer lugar do mundo), a partir de Curitiba (Brasil). Sou formado em Ciência da Computação pela Universidade Federal do Paraná (UFPR) e atuo transformando desafios complexos de negócio em soluções de IA escaláveis e prontas para produção.

I'm a Senior Data Scientist and AI/ML Engineer working 100% remotely (for clients anywhere in the world), from Curitiba, Brazil. I hold a B.Sc. in Computer Science from the Federal University of Parana (UFPR) and specialize in transforming complex business challenges into scalable, production-ready AI solutions.

Já passei por e-commerce, logística, fintech, seguros, varejo, bancos, saúde e consultoria, sempre com foco em ownership end-to-end: da definição do problema e análise exploratória até o desenvolvimento, deploy, monitoramento dos modelos e automação de processos.

Over my career, I've worked across e-commerce, logistics, fintech, insurance, retail, banking, healthcare and consulting, always with a focus on end-to-end ownership: from problem definition and data exploration through model development, deployment, monitoring, and process automation.

Também atuo como educador e autor. Escrevi o livro "Python para Análise de Dados: Do Zero ao Insight" e ensino Python para profissionais em transição para a área tech. Acredito que as melhores soluções surgem quando pensamento criativo e execução técnica rigorosa andam juntos.

I'm also an educator and author. I wrote "Python para Analise de Dados: Do Zero ao Insight" and teach Python to professionals transitioning into tech. I believe the best solutions come from combining creative thinking with rigorous technical execution.

Data Science AI/ML GenAI Data Engineering Process Automation Analytics Big Data
"O que me move não é o que é novo, mas o que ainda não foi imaginado." "What moves me is not what is new, but what is not yet imagined."

Impacto no Negócio Business Impact

Meço cada modelo pelo valor que ele gera no negócio, não apenas pela acurácia. Every model I build is measured by the business value it delivers, not just accuracy scores.

Ownership End-to-End End-to-End Ownership

Do dado bruto à API em produção, cuido do ciclo inteiro para entregar uma solução que funciona de ponta a ponta. From raw data to production API. I handle the full lifecycle so you get a complete solution.

Comunicação Clara Clear Communication

Traduzo temas técnicos complexos em insights acionáveis para qualquer público. I translate complex technical concepts into actionable insights for any audience.

Resultados Comprovados Proven Results

6.000x de aceleração em pipeline. 33% a mais de acurácia. 13% de redução de custos. Números reais, impacto real. 6,000x pipeline speedup. 33% accuracy improvement. 13% cost reduction. Real numbers, real impact.

6.000x Ganho de Performance em Pipelines Pipeline Performance Gain
33% Aumento de Acurácia em Previsões Prediction Accuracy Boost
13% Redução de Custos Operacionais Operational Cost Reduction
22% Aumento na Satisfação do Cliente Customer Satisfaction Increase
Trajetória Career Path

Experiência Profissional Professional Experience

Fundador & AI/ML Engineer Founder & AI/ML Engineer

Facilia (Consultoria em Data Science e IA) Facilia (Data Science & AI Consulting)
Nov 2025 - Presente Nov 2025 - Present
Consultoria, Data Science e IA Consulting, Data Science & AI
  • Fundei a Facilia, minha consultoria de Data Science e IA. Entre os projetos atuais, atuo como PJ na triggo.ai, entregando soluções end-to-end para clientes enterprise.
  • Founded Facilia, my Data Science & AI consulting practice; among current engagements, providing services as PJ to triggo.ai, delivering end-to-end projects to enterprise clients.
  • Funcional Health: processo e rotina automatizada (com data warehouse) para obtenção, limpeza e armazenamento de dados do CNES/DataSUS sobre estabelecimentos de saúde. Construí um crawler de web scraping controlado por LLM para enriquecer informações sobre 200 mil+ estabelecimentos.
  • Funcional Health: end-to-end process automation for ingesting, cleaning and storing CNES/DataSUS healthcare data into a data-warehouse-like layer; built an LLM-controlled web scraping crawler enriching 200K+ healthcare establishments.
  • Guima Conseco: construí a camada Gold dimensional do data warehouse, criei e implementei KPIs e indicadores, montei dashboards para apresentação dessas métricas e desenvolvi modelos preditivos de ML para absenteísmo e turnover.
  • Guima Conseco: designed the dimensional Gold layer of the data warehouse, defined KPIs and operational metrics, built executive dashboards, and delivered predictive ML models for absenteeism and turnover.
  • Natura: refatorei o módulo de forecast usado no sistema de otimização do catálogo de produtos.
  • Natura: refactored the forecasting module powering their product catalog optimization system.

Instrutor de Python e Autor Python Instructor & Author

Do Zero ao Insight
Ago 2024 - Presente Aug 2024 - Present
Educação e Treinamento Education & Training
  • Criei e ministro um curso prático de Python para Análise de Dados baseado em projetos reais.
  • Designed and delivered a hands-on Python for Data Analysis course based on real-world data projects.
  • Autor do livro "Python para Análise de Dados: Do Zero ao Insight", utilizado como material principal do curso.
  • Authored the book "Python para Análise de Dados: Do Zero ao Insight", used as primary course material.
  • Mentoro públicos diversos em transição de carreiras de negócios, engenharia e ciências para a área de dados.
  • Mentor diverse audiences moving from business, engineering and science backgrounds into data roles.

Senior Data Scientist

ALLOS
Mai 2024 - Jan 2025 May 2024 - Jan 2025
Retail Real Estate e Operadora de Shopping Centers Retail Real Estate & Shopping Mall Operator
  • Liderei iniciativas de Big Data e analytics para apoiar as operações dos shoppings e a estratégia do programa de fidelidade.
  • Led Big Data and analytics initiatives supporting mall operations and loyalty program strategy.
  • Rearquitetei pipelines de Big Data com PySpark e SQL, reduzindo o tempo de processamento end-to-end em 6.000x.
  • Re-architected Big Data pipelines with PySpark and SQL, reducing end-to-end processing time by 6,000x.
  • Desenvolvi um algoritmo de fuzzy matching com 98% de acurácia para mapear lojas e benefícios do programa de fidelidade, automatizando um fluxo que antes era feito manualmente.
  • Built a fuzzy matching algorithm with 98% accuracy for loyalty program store-benefit mapping, automating a previously manual curation flow.
  • Entreguei dashboards de KPIs no Databricks, viabilizando decisões executivas mais rápidas e assertivas.
  • Delivered KPI dashboards in Databricks, enabling faster and more informed executive decision-making.

Senior Data Scientist

Prudential do Brasil
Fev 2024 - Mai 2024 Feb 2024 - May 2024
Seguros e Serviços Financeiros Insurance & Financial Services
  • Estruturei do zero a área de IA da empresa, definindo o roadmap de Inteligência Artificial.
  • Launched the company's AI practice from the ground up, defining the AI roadmap.
  • Liderei projetos pioneiros de GenAI e analytics avançado.
  • Led foundational GenAI and advanced analytics projects.
  • Fortaleci a governança e a qualidade dos dados, evoluindo o data lake para suportar analytics crítico ao negócio.
  • Strengthened data governance and quality, enhancing the data lake to support business-critical analytics.

Principal Data Scientist (anteriormente Lead) Principal Data Scientist (prev. Lead)

Swap (Fintech)
Jan 2022 - Jun 2023
Fintech e Pagamentos Digitais Fintech & Digital Payments
  • Co-fundei o time de Data Science e fui promovido de Lead a Principal, alinhando a estratégia executiva à execução técnica.
  • Co-founded the Data Science team and was promoted from Lead to Principal, aligning executive strategy with technical execution.
  • Desenvolvi algoritmos que garantiram 100% de rastreabilidade e reconciliação das transações financeiras, atendendo às exigências de compliance do Banco Central do Brasil.
  • Developed algorithms ensuring 100% traceability and reconciliation of financial transactions, meeting Central Bank of Brazil compliance requirements.
  • Implementei uma infraestrutura de dados em nível regulatório cobrindo todo o ciclo de vida do dado, requisito-chave para a certificação como Instituição de Pagamento.
  • Implemented regulatory-grade data infrastructure covering the full data lifecycle, a key requirement for Payment Institution certification.
  • Automatizei a reconciliação de transações entre múltiplos processadores e bancos, eliminando um grande gargalo operacional que antes era feito na mão.
  • Automated transaction reconciliation across multiple processors and banks, removing a major manual operational burden.
  • Estruturei metodologias ágeis orientadas a resultado no time de DS e atuei como mentor dos demais membros.
  • Structured agile, results-oriented methodologies for the DS team and mentored team members.

Data Scientist

Delivery Center
Mai 2020 - Dez 2021 May 2020 - Dec 2021
Logística e Last-Mile Delivery Logistics & Last-Mile Delivery
  • Otimizei rotas de entrega com ML e algoritmos bio-inspirados, reduzindo custos operacionais em 13% e elevando a satisfação do cliente em 22%.
  • Optimized delivery routes with ML and bio-inspired algorithms, cutting operational costs by 13% and raising customer satisfaction by 22%.
  • Construí um sistema de predição de risco em tempo real que possibilitou intervenções proativas para evitar falhas nas entregas.
  • Built a real-time risk prediction system enabling proactive interventions to prevent delivery failures.
  • Entreguei modelos de previsão de demanda (séries temporais) e de precificação margin-aware, gerando 7% de crescimento de receita.
  • Delivered time-series demand forecasting and margin-aware pricing models, driving 7% revenue growth.

Junior Data Scientist

Olist
Jul 2018 - Abr 2020 Jul 2018 - Apr 2020
E-commerce e Marketplace E-commerce & Marketplace
  • Desenvolvi um estimador de prazo de entrega baseado em ML, aumentando a acurácia em 33% e reduzindo reclamações em 19%.
  • Developed an ML-based delivery estimator, improving prediction accuracy by 33% and reducing customer complaints by 19%.
  • Implementei o onboarding automatizado de clientes, reduzindo o tempo total em 67% (de três semanas para uma).
  • Implemented automated customer onboarding, cutting completion time by 67% (from three weeks to one week).
  • Construí um sistema de categorização de produtos com NLP, reduzindo o trabalho manual em 65% e elevando a qualidade do catálogo.
  • Built an NLP product categorization system, reducing manual input by 65% and improving catalog quality.

Consultor Oracle Retail Oracle Retail Consultant

Logic Information Systems
Mar 2018 - Jun 2018 Mar 2018 - Jun 2018
Consultoria em Tecnologia de Varejo Retail Technology Consulting
  • Apoiei o planejamento de vendas e forecasting de grandes varejistas analisando dados históricos e alinhando insights aos objetivos de negócio, com Oracle RPAS.
  • Supported sales planning and forecasting for major retailers by analyzing historical data and aligning insights with business goals, using Oracle RPAS.
  • Implementei automação de processos (RPA) no Backoffice, reduzindo tempo de execução e erros operacionais.
  • Implemented process automation (RPA) in the Backoffice department, reducing execution time and operational errors.

Estagiário (Desenvolvedor VBA e RPA) Intern (VBA & RPA Developer)

HSBC
Jan 2013 - Jun 2014
Bancos e Fundos de Investimento Banking & Investment Funds
  • Desenvolvi e mantive soluções em VBA para apoiar as operações do dia a dia do departamento de fundos de investimento.
  • Developed and maintained VBA-based solutions supporting day-to-day operations of the investment funds department.
  • Implementei automação de processos (RPA) em fluxos ligados aos fundos, reduzindo bastante o tempo de execução e os erros operacionais.
  • Implemented process automation (RPA) for fund-related workflows, significantly reducing execution time and operational errors.
  • Contribuí para o Plano de Continuidade de Negócios (BCP) do departamento, apoiando iniciativas de mitigação de riscos.
  • Contributed to the department's Business Continuity Plan (BCP), supporting risk mitigation initiatives.
Portfólio Portfolio

Cases em Destaque Flagship Case Studies

Quatro projetos que resumem o tipo de impacto que entrego. Four projects that capture the kind of impact I deliver.

6.000x mais rápido faster
Data Engineering · ALLOS

Pipelines de Big Data 6.000x mais rápidos Big Data Pipelines, 6,000× Faster

Problema: jobs em batch que rodavam por horas atrasavam decisões executivas em uma operadora de shoppings.

Problem: hours-long batch jobs were delaying executive decisions at a shopping-mall operator.

Ação: rearquitetei o pipeline em PySpark + Delta Lake, com particionamento e cache estratégico, transformando processamento em near-real-time.

Action: re-architected the pipeline in PySpark + Delta Lake with strategic partitioning and caching, turning batch into near real-time.

Resultado: 6.000× de speedup end-to-end e dashboards executivos com dados frescos.

Result: 6,000× end-to-end speedup and fresh data on executive dashboards.

PySparkDatabricksDelta LakeAirflow
-13% custo · +22% satisfação cost · +22% satisfaction
Optimization · Delivery Center

Otimização de Rotas Last-Mile Last-Mile Route Optimization

Problema: custos de entrega altos e janelas estouradas em operação last-mile.

Problem: high delivery costs and missed time windows in a last-mile operation.

Ação: uni predições de ML a algoritmos bio-inspirados (genéticos + colônia de formigas) considerando tráfego, capacidade e janelas em tempo real.

Action: combined ML predictions with bio-inspired algorithms (genetic + ant colony) factoring in traffic, capacity and time windows in real time.

Resultado: 13% de redução de custo operacional e 22% a mais de satisfação do cliente.

Result: 13% operational cost reduction and 22% higher customer satisfaction.

PythonOR-ToolsGenetic AlgorithmsFastAPI
-40% tickets de suporte support tickets
Generative AI · Enterprise

Assistente de Conhecimento com RAG RAG-Powered Knowledge Assistant

Problema: 50.000+ documentos internos eram inacessíveis e geravam volume alto de tickets de suporte.

Problem: 50,000+ internal docs were hard to search and drove a heavy support-ticket load.

Ação: arquitetei um sistema RAG com LangChain, FAISS e GPT-4/Claude que devolve respostas instantâneas com fontes citadas.

Action: built a RAG system with LangChain, FAISS and GPT-4/Claude that returns instant, sourced answers.

Resultado: 40% menos tickets de suporte e tempo de resposta caindo de horas para segundos.

Result: 40% fewer support tickets and response time down from hours to seconds.

LangChainFAISSOpenAIFastAPI
200K+ estabelecimentos enriquecidos facilities enriched
Automation · Funcional Health

Pipeline de Saúde Orquestrado por LLM LLM-Orchestrated Healthcare Pipeline

Problema: a curadoria de dados de estabelecimentos de saúde do CNES/DataSUS era manual e não escalava.

Problem: manual curation of CNES/DataSUS healthcare data didn't scale.

Ação: construí um pipeline end-to-end com um crawler de scraping controlado por LLM que enriquece e valida dados sem regras frágeis.

Action: built an end-to-end pipeline with an LLM-controlled scraping crawler that enriches and validates data without brittle rules.

Resultado: 200 mil+ estabelecimentos enriquecidos automaticamente, eliminando o trabalho manual.

Result: 200K+ facilities enriched automatically, eliminating manual work.

LangChainPlaywrightAirflowPostgreSQL

Mais Projetos More Projects

Filtre por categoria para explorar o portfólio completo. Filter by category to explore the full portfolio.

Generative AI & LLMs

RAG-Powered Enterprise Knowledge Assistant

Built an intelligent Q&A system using Retrieval-Augmented Generation (RAG) over 50,000+ internal documents. Integrated LangChain, FAISS vector store, and GPT-4/Claude APIs to provide instant, sourced answers to employee queries, reducing support ticket volume by 40%.

LangChainFAISSFastAPIPythonDocker
40% reduction in support tickets
Generative AI & LLMs

AI-Powered Insurance Policy Analyzer

Developed a GenAI system for Prudential that automatically analyzes insurance policies, extracts key terms, and generates risk summaries. Used LLMs with structured output parsing and prompt engineering to automate what previously required hours of manual review per policy.

LangChainOpenAIPydanticStreamlitPostgreSQL
85% faster policy analysis
Generative AI & LLMs

Multi-Agent AI Workflow for Market Research

Designed a multi-agent system using CrewAI and LangGraph where specialized AI agents collaborate to perform market research: one scrapes data, another analyzes competitors, and a third generates executive reports. Reduced research cycle from 2 weeks to 2 days.

CrewAILangGraphPythonQdrantGradio
85% faster research cycle
Generative AI & LLMs

Custom LLM Fine-Tuning for Financial Compliance

Fine-tuned open-source LLMs (LLaMA, Mistral) on financial regulation datasets for a fintech client. The specialized model automatically classifies transactions, flags compliance issues, and generates audit-ready explanations, achieving 96% accuracy on regulatory queries.

Hugging FaceLoRAPyTorchAWS SageMakerMLflow
96% regulatory query accuracy
Machine Learning

ML-Based Delivery Time Estimator

Developed an ensemble ML model (XGBoost + LightGBM) at Olist that predicted delivery times with 33% more accuracy than the previous rule-based system. Integrated geospatial features, carrier performance data, and seasonal patterns. Reduced customer complaints by 19%.

XGBoostLightGBMScikit-learnFlaskDocker
33% accuracy improvement, -19% complaints
Machine Learning

Real-Time Delivery Risk Prediction System

Built a real-time risk scoring engine at Delivery Center that monitors active deliveries and predicts failure probability. The system triggers proactive interventions (driver reassignment, customer notifications) when risk thresholds are exceeded, preventing delivery failures before they happen.

PythonScikit-learnFastAPIRedisPostgreSQL
Proactive failure prevention
Machine Learning

Time-Series Demand Forecasting & Pricing Engine

Created a demand forecasting system combining ARIMA, Prophet, and gradient boosting models at Delivery Center. Paired with a margin-aware optimization layer that dynamically adjusts pricing based on demand, capacity, and cost signals, driving 7% revenue growth.

ProphetXGBoostOptunaAirflowDatabricks
7% revenue growth
Machine Learning

Customer Churn Prediction & Retention System

Developed a churn prediction model for a retail client that identifies at-risk customers 30 days before they churn. Combined behavioral features, RFM analysis, and survival models to generate risk scores and personalized retention strategies, reducing churn by 18%.

LightGBMScikit-learnPySparkAirflowStreamlit
18% churn reduction
Machine Learning

Genetic Programming for Cancer Prediction

Capstone research project at UFPR developing a novel genetic programming-based clustering algorithm for gene expression analysis. The algorithm discovers patterns in high-dimensional biological data to assist in cancer type classification and prediction.

PythonEvolutionary AlgorithmsNumPyScikit-learn
Novel research contribution
Machine Learning

Absenteeism & Turnover Prediction Models

Built predictive ML models (via Facilia / triggo.ai, client: Guima Conseco) to anticipate employee absenteeism and voluntary turnover. Combined HR data, operational signals, and behavioral features to generate risk scores that feed into retention and workforce-planning workflows.

PythonLightGBMScikit-learnPandasDatabricks
Proactive HR decision-making
NLP & Text Analytics

NLP Product Categorization Engine

Built an NLP-based automatic product categorization system at Olist that analyzes product titles, descriptions, and attributes to classify items into the correct taxonomy. Reduced manual input by 65% and improved catalog quality, directly impacting customer satisfaction.

spaCyTF-IDFScikit-learnFastAPIPostgreSQL
65% reduction in manual classification
NLP & Text Analytics

Customer Review Sentiment Analysis Pipeline

Developed an end-to-end sentiment analysis pipeline that processes thousands of customer reviews daily, extracting sentiment, key topics, and actionable insights. Used transformer-based models (BERT) fine-tuned on domain-specific data to achieve 92% accuracy on Portuguese text.

Hugging FaceBERTPyTorchAirflowStreamlit
92% sentiment accuracy
NLP & Text Analytics

Intelligent Document Processing (IDP) System

Built an IDP system that extracts structured data from unstructured documents (invoices, contracts, reports) using OCR + NLP + LLMs. The system handles multiple document formats, validates extracted fields, and integrates with client ERP systems via API.

TesseractLangChainspaCyFastAPIDocker
90% automation rate
Web Scraping & Data Extraction

LLM-Powered Intelligent Web Scraper

Developed an adaptive web scraping system that uses LLMs to understand page structure and extract data without brittle CSS selectors. The AI agent navigates dynamic pages, handles pagination, CAPTCHAs, and anti-bot measures, and outputs clean structured data in any schema.

LangChainPlaywrightOpenAIPythonMongoDB
95% extraction accuracy on dynamic sites
Web Scraping & Data Extraction

E-commerce Competitive Price Monitor

Built a scalable price monitoring system that scrapes 100,000+ products daily across major e-commerce platforms. Features include price change alerts, historical price tracking, competitor analysis dashboards, and automated repricing recommendations based on market positioning.

ScrapySeleniumPostgreSQLAirflowStreamlit
100K+ products monitored daily
Web Scraping & Data Extraction

Real Estate Market Intelligence Platform

Created an automated data collection platform that scrapes property listings, rental prices, and market trends from multiple real estate portals. Includes geocoding, deduplication with fuzzy matching, and a dashboard for market analysis and investment decision support.

ScrapyBeautifulSoupPostgreSQLPlotlyDocker
500K+ listings tracked
Web Scraping & Data Extraction

B2B Lead Generation & Enrichment Engine

Built a lead generation system that scrapes business directories, LinkedIn profiles, and company websites to build enriched prospect databases. Uses NLP to classify industry, company size, and technologies used, enabling highly targeted outreach for sales teams.

PlaywrightspaCyPythonMongoDBFastAPI
10K+ enriched leads/month
Data Engineering

Big Data Pipeline Optimization (6,000x Speedup)

Re-architected and optimized data pipelines at ALLOS using PySpark, SQL, and ML techniques. Transformed batch processing jobs that took hours into near-real-time pipelines, achieving a 6,000x performance improvement. Enabled data freshness for executive dashboards.

PySparkDatabricksSQLDelta LakeAirflow
6,000x processing speedup
Data Engineering

Regulatory-Grade Data Infrastructure (Fintech)

Designed and implemented the complete data infrastructure at Swap covering ingestion, transformation, storage, and governance. Built to Central Bank of Brazil standards with 100% transaction traceability, audit trails, and data quality monitoring, critical for Payment Institution certification.

PythonAirflowPostgreSQLDockerAWS
100% compliance achieved
Data Engineering

Real-Time ETL Pipeline with Event Streaming

Built a real-time data ingestion and transformation pipeline that processes millions of events daily from multiple sources. Uses event-driven architecture with message queues for reliable delivery, schema evolution support, and automated data quality checks.

Apache KafkaPySparkAirflowPostgreSQLDocker
Millions of events/day processed
Data Engineering

Dimensional Data Warehouse (Gold Layer)

Designed and implemented the dimensional Gold layer of the data warehouse (via Facilia / triggo.ai, client: Guima Conseco). Modeled facts and dimensions aligned with business processes, powering KPIs, operational metrics, and executive dashboards, and feeding downstream predictive models.

SQLPySparkDatabricksStar SchemaAirflow
Unified business metrics layer
Process Automation

Automated Customer Onboarding System

Designed and implemented an intelligent onboarding automation system at Olist that streamlined the seller registration process. Automated document validation, data verification, and account setup workflows, reducing completion time by 67% (from 3 weeks to 1 week).

PythonAirflowPostgreSQLREST APIsDocker
67% faster onboarding
Process Automation

Financial Report Generation Automation

Built an end-to-end automated reporting system that pulls data from multiple sources, runs validation checks, applies business rules, and generates formatted financial reports. Includes anomaly detection that flags unusual values for human review before distribution.

PythonPandasAirflowJinja2Slack API
90% reduction in manual reporting
Process Automation

Transaction Reconciliation Automation

Developed an automated reconciliation system at Swap that matches financial transactions across multiple payment processors, banks, and internal systems. Handles edge cases like partial payments, refunds, and chargebacks with 100% traceability for audit compliance.

PythonPySparkPostgreSQLAirflowFastAPI
100% transaction traceability
Process Automation

LLM-Driven Healthcare Data Pipeline

Designed an end-to-end automated pipeline (via Facilia / triggo.ai, client: Funcional Health) for ingesting, cleaning, and storing healthcare establishment data from CNES/DataSUS into a data-warehouse-like layer. An LLM-controlled web scraping crawler enriches information on 200K+ establishments, replacing a previously manual curation process.

PythonLangChainPlaywrightAirflowPostgreSQL
200K+ establishments automated
Process Automation

Investment Funds Workflow RPA (VBA + RPA)

Automated day-to-day operational workflows in HSBC's investment funds department using VBA-based solutions and Robotic Process Automation (RPA). Significantly reduced execution time and operational errors in fund-related processes, while supporting the department's Business Continuity Plan (BCP).

VBARPAExcelMacros
Reduced execution time & errors
Process Automation

Retail Backoffice RPA Automation

Implemented Robotic Process Automation (RPA) routines in the Backoffice department of a major retail consulting client at Logic Information Systems. The automation reduced execution time and operational errors in planning and forecasting workflows powered by Oracle RPAS.

RPAOracle RPASVBASQL
Faster backoffice, fewer errors
Optimization

Last-Mile Delivery Route Optimization

Built an intelligent route optimization engine at Delivery Center combining ML predictions with bio-inspired algorithms (genetic algorithms, ant colony optimization). The system considers real-time traffic, driver capacity, delivery windows, and cost constraints to generate optimal routes.

PythonOR-ToolsGenetic AlgorithmsFastAPIRedis
13% cost reduction, +22% satisfaction
Optimization

Dynamic Pricing & Margin Optimization

Created a margin-aware pricing optimization system that balances revenue maximization with competitive positioning. Uses demand elasticity models, competitor pricing signals, and cost structures to recommend optimal price points in real-time, contributing to 7% revenue growth.

PythonOptunaXGBoostFastAPIDatabricks
7% revenue increase
Analytics & BI

Executive KPI Dashboard Suite

Designed and implemented a comprehensive KPI dashboard suite in Databricks at ALLOS covering sales performance, customer engagement, and operational efficiency. Features drill-down capabilities, automated alerts, and natural language summaries for C-level stakeholders.

DatabricksSQLPlotlyPythonPySpark
Data-driven executive decisions
Analytics & BI

Fuzzy Matching for Loyalty Program Mapping

Built a fuzzy matching algorithm at ALLOS that maps stores to benefits in a loyalty program with 98% accuracy. Handles name variations, abbreviations, and typos across thousands of store entries, enabling automated benefit assignment that previously required manual curation.

PythonFuzzyWuzzyPySparkDatabricksSQL
98% matching accuracy
Analytics & BI

Data Lake Architecture & Governance Framework

Architected and implemented a data lake at Prudential do Brasil with proper governance layers: data cataloging, quality monitoring, access controls, and lineage tracking. Enabled the organization to transition from siloed spreadsheets to a unified, trusted data platform.

AWS S3GlueAthenaPythonAirflow
Unified data platform established
Stack Técnico Tech Stack

Competências e Tecnologias Skills & Technologies

Machine Learning & AI

Supervised Learning Unsupervised Learning Deep Learning NLP Generative AI LLMs / RAG Fine-Tuning Time Series Evolutionary Computation

Frameworks & Libraries

TensorFlow PyTorch Scikit-learn XGBoost LightGBM Hugging Face LangChain LangGraph CrewAI LlamaIndex Optuna

Programação e Dados Programming & Data

Python SQL PySpark Pandas NumPy Databricks

Data Engineering

ETL/ELT Pipelines Apache Airflow Apache Spark PostgreSQL MySQL MongoDB FAISS Pinecone Chroma Qdrant

Cloud & MLOps

AWS GCP SageMaker Kubeflow MLflow FastAPI Flask Docker Streamlit Gradio

Automação de Processos Process Automation

Process Automation RPA VBA Oracle RPAS Workflow Orchestration LLM-Driven Automation Web Scraping

Visualização de Dados Visualization

Matplotlib Seaborn Plotly KPI Dashboards Data Storytelling
Formação Education

Formação Acadêmica Academic Background

Bacharelado em Ciência da Computação B.Sc. in Computer Science

Universidade Federal do Paraná (UFPR), Brasil Federal University of Parana (UFPR), Brazil 2017

TCC: algoritmo de clustering baseado em programação genética para análise de expressão gênica e predição de câncer.

Capstone Project: Genetic programming-based clustering algorithm for gene expression analysis and cancer prediction.

Iniciação Científica Undergraduate Research (Iniciação Científica)

NR2 - Núcleo de Redes Sem Fio e Redes Avançadas, UFPR NR2 - Wireless & Advanced Networks Lab, UFPR 2012

Projeto: SIMTUR (Sistema Inteligente de Monitoramento de Tráfego Urbano), focado em soluções integradas de mobilidade para rastreamento, comunicação, controle e monitoramento do transporte urbano e do tráfego.

Project: SIMTUR (Intelligent Urban Traffic Monitoring System), modeling integrated mobility solutions for tracking, communication, control, and monitoring of urban transport and traffic.

Contribuição: analisei dados experimentais de testes do protocolo MPOLSR em Vehicular Ad Hoc Networks (VANETs) e apliquei um algoritmo bio-inspirado baseado em fluxo para investigar ganhos de performance e eficiência da rede.

Contribution: Analyzed experimental data from MPOLSR protocol tests in Vehicular Ad Hoc Networks (VANETs) and applied a bio-inspired stream-based algorithm to investigate improvements in network performance and efficiency.

Autor Publicado Published Author

"Python para Análise de Dados: Do Zero ao Insight" 2024

Autor de um livro completo sobre Python para Análise de Dados, usado como material principal do curso. Vai dos fundamentos de Python até projetos reais de análise de dados.

Authored a comprehensive Python for Data Analysis book used as primary course material. Covers everything from Python fundamentals to real-world data analysis projects.

Entre em Contato Get in Touch

Vamos Criar Algo
Incrível Juntos Let's Build Something
Amazing Together

Estou aberto a projetos freelance, consultorias e colaborações de longo prazo. Se você precisa de uma solução de IA, um pipeline de dados ou um modelo de ML, vamos conversar. I'm available for freelance projects, consulting, and long-term collaborations. Whether you need an AI solution, data pipeline, or ML model, let's talk.

Email

horning.walter@gmail.com

LinkedIn

walter-horning-junior

Telefone & WhatsApp Phone & WhatsApp

+55 41 9 9849 7081