Andre Emidio

Professional with more than 14 years of experience in technology, with the last 6 years focused on Artificial Intelligence engineering, Computer Vision, Machine Learning, and MLOps. Throughout my career, I have developed and deployed solutions in sectors such as the food industry, manufacturing, agribusiness, education, and financial services, with active participation in projects ranging from the design of data architectures to the deployment of models in large-scale production environments. My background includes the creation of complete predictive maintenance platforms with expansion into the Latin American market, high-precision optical recognition systems, computer vision pipelines for industries with connectivity restrictions, and MLOps infrastructures for financial institutions.

I work with tools and technologies such as Python, C++, OpenCV, PyTorch, TensorFlow, YOLO, XGBoost, SHAP, MLflow, as well as cloud architectures with MS Azure and AWS, container orchestration with Docker and Kubernetes, industrial protocols such as OPC UA, and system integration through REST APIs, Message Queues, and IoT Edge. I hold a degree in Computer Network Technology from ETEP and was a special student with maximum performance in the Master's Program in Computer Science at the Federal University of Santa Catarina, with certifications in deep learning, MLOps, and time series.

Education

Special Student in the Master's Program in Computer Science | UFSC - Federal University of Santa Catarina
Completed: 2022

Computer Network Technology | ETEP Faculdades
Completed: 2015

Results & Recognition

International-Scale Cost Avoidance Generation | T3 Labs AI

Development and deployment of a predictive maintenance platform in industrial plants in Latin America, generating approximately USD 362 thousand in cost avoidance in a pilot project through early failure detection and intervention planning before unplanned downtime.

Industrial AI Platform Developed from Scratch | T3 Labs AI

Full design and delivery of a predictive maintenance platform to replace a consolidated legacy system in the market, including the development of an autonomous training system that eliminated dependency on technical intervention for model updates and later expansion to multiple plants and process areas in Latin America.

Model Deployment Infrastructure Modernization | NTT Data

Evolution of the ModelHub platform through the implementation of serverless architectures on AWS, CI/CD pipelines with GitHub Actions, and test coverage above 90% using Pytest/BDD, bringing the platform experience closer to consolidated market solutions and reducing bureaucracy in the model deployment process for data scientists.

Modernization of Lottery Card Reading Library | INDT

Design and modernization of an OMR reading library for lottery cards, including legacy system analysis, portability to Debian-based Linux environments, and development of components in C and C++, ensuring compatibility with updated infrastructures and technological evolution of the solution.

Computer Vision in a Highly Restricted Industrial Environment | Engineering Brasil

Development of an automatic poultry counting system in an industrial slaughterhouse using edge computing, C++, and the Hungarian optimization algorithm, operating in an environment with severe connectivity restrictions and enabling automated measurement of the industrial plant's daily production.

Data Aggregation Product Adopted by Financial Institutions | Datatype

Development of a platform for consolidating CPF, chassis, and vehicle data, adopted by companies such as Suhai Seguradora and C6 Bank for risk analysis, client identification, and asset valuation processes, with a backend architecture built entirely from scratch.

Professional Experience

Machine Learning, Computer Vision & Predictive Analytics Consulting | T3 Labs AI

Role: Co-Founder & Machine Learning Engineering Consultant 09/2023 - Present

Served as Co-Founder and technical lead responsible for the design, development, and expansion of a predictive maintenance platform based on intelligent agents for the industrial sector, with deployments in plants in Brazil and Latin America.

• Full development of a predictive maintenance platform with boolean agents and Machine Learning models based on XGBoost with balancing via SMOTE, integrated with Azure services such as Machine Learning, Event Hubs, Azure Data Explorer (KustoDB), Container Apps, IoT Hub, IoT Edge, and Azure OpenAI, with CI/CD pipelines for deployment automation and migration from Azure Functions to containers with Celery and Celery Beat for greater operational stability.
• Integration of industrial systems through the OPC UA protocol with OPC Publisher on IoT Edge, connected to industrial data historians, with development of tools in Python and AsyncUA to automate the collection of sensor NodeIDs, enabling real-time ingestion of operational data into the Azure cloud.
• Implementation of SHAP (Shapley Values) for model prediction explainability, with contextualized alerts sent via Microsoft Teams, email, and WhatsApp, allowing maintenance teams to accurately identify the sensors and variables responsible for each detected anomaly.
• Development of an RCA (Root Cause Analysis) tool based on RAG with Azure OpenAI, integrating data from SQL Server, SharePoint, PDFs, and work instructions, with the goal of accelerating the creation of root cause reports and reducing the time teams spend manually searching for information.
• Creation of integrations for data collection from specialized industrial suppliers, development of APIs for integrating industrial historians using the OPC UA protocol into internal monitoring platforms, and construction of environmental monitoring solutions for industrial processes with automatic notification delivery.
• Conducted training for new users in Portuguese and Spanish, participated in meetings with sponsors and architecture teams in a multilingual environment, and provided continuous support to the operational teams of the plants served.

Machine Learning Engineering & MLOps | NTT Data

Role: Machine Learning Engineer 10/2024 - 03/2026

Worked on the evolution of a corporate Machine Learning model deployment platform for a large Brazilian bank, with responsibility for the full implementation cycle, software quality, cloud infrastructure development, and technical support for data science teams.

• Developed and evolved the ModelHub platform, proposing and implementing architectural improvements that reduced bureaucracy in the model deployment process and brought the platform experience closer to consolidated market solutions such as Amazon SageMaker.
• Built serverless architectures for AI model deployment using AWS Lambda and Amazon ECS, structured infrastructure as code with Terraform, and automated CI/CD pipelines with GitHub Actions, ensuring reproducible and scalable deployments.
• Developed internal libraries to standardize processes and increase the productivity of the squad and partner areas, creating Quick Commands in StackSpot AI to automate recurring tasks such as technical documentation generation and unit test creation through prompt engineering.
• Implemented automated tests with Pytest using a BDD approach, achieving coverage above 90% and increasing application reliability, with complementary work on bug fixing, code refactoring, and the application of SOLID and Clean Architecture principles.
• Developed integrations with REST APIs and PostgreSQL databases, in addition to creating Proofs of Concept to validate new technologies, with active contribution to discussions about the adoption of AI agents in the platform.
• Provided technical mentoring to data scientists and new team members through pair programming, onboarding, and dissemination of software engineering best practices, working in collaboration with operations squads, Itau Asset Management, and business teams.

Senior Software Engineering, OpenCV & OMR | Compass UOL / Poliedro Educacao

Role: Senior Software Engineer 10/2023 - 09/2024

Architected and delivered an automated Optical Mark Recognition (OMR) system for a high-volume educational institution, leading image processing research and technical team management.

• Engineered a high-accuracy Optical Mark Recognition (OMR) system for automated grading of exams and mock tests, including K-12 and elite military preparation courses such as ITA/IME, achieving approximately 98% accuracy across processed documents.
• Designed and implemented a scalable distributed architecture using Apache Pulsar for image processing across workers, with deployment on Azure using Kubernetes (AKS) and automated CI/CD pipelines through Azure DevOps.
• Led research and experimentation in Digital Image Processing (DIP), evaluating detection and filtering techniques for marker recognition and documenting architectural decisions and performance benchmarks in technical reports.
• Developed new features and improved system stability throughout the software lifecycle, integrating complex image processing workflows into message queue systems and maintaining technical alignment with stakeholders.
• Conducted architectural training and technical deep dives on processing flows for the engineering team, supporting autonomy and long-term operational continuity of the platform.

Research, Development & Computer Vision | INDT

Role: Senior Computer Vision Engineer 01/2023 - 09/2023

Worked on an R&D project for client Diebold Nixdorf, focused on modernizing a legacy software library, developing components in C and C++, and leading applied research and technical documentation initiatives to meet SUFRAMA requirements.

• Modernized a legacy library originally developed in 2005, analyzing the existing architecture, identifying obsolete dependencies, and porting the solution to Debian 9-based environments, ensuring compatibility and operational continuity in the client's updated infrastructures.
• Developed and ported libraries in C++ and C to support project functionality, defining and prioritizing technical tasks and suggesting integration of the solution with other existing systems in the client's architecture.
• Conducted Proofs of Concept to validate new ideas and explore technological opportunities, with potential evolution into future products in the organization's solution portfolio.
• Wrote detailed technical reports on research and development activities to meet the regulatory requirements of the Superintendence of the Manaus Free Trade Zone (SUFRAMA), ensuring formal records of technological innovation initiatives.
• Provided technical mentoring to a mid-level developer with no previous experience in computer vision, presenting fundamental concepts, monitoring implementations, and supporting the resolution of technical challenges, accelerating the professional's autonomy in the project.

Artificial Intelligence Engineering | Syngenta Digital

Role: AI Engineer 04/2022 - 09/2022

Worked on the development and evolution of a remote imaging platform for processing and distributing satellite images, with contributions in backend development, system integration, applied research, and cloud infrastructure optimization.

• Developed services and internal tools to support the remote imaging platform, implementing integrations through REST APIs between systems and database modeling using PostgreSQL, ensuring interoperability among the platform's applications.
• Maintained and evolved the existing image platform in parallel with the development of the new version, identifying and fixing technical issues and infrastructure adjustments to ensure operational continuity during the transition.
• Developed solutions using Amazon Web Services, working on infrastructure configuration with a focus on scalability, reliability, and optimization of operating costs.
• Worked on an agricultural scouting tool focused on collecting and analyzing field plot data, contributing to the integration of geospatial data and supporting the development of decision-oriented applications for agribusiness.
• Investigated tools and approaches to optimize image processing, creating software delivery pipelines and contributing to the continuous evolution of the platform services' quality and efficiency.

Applied Computer Vision & Edge Computing | Engineering Brasil

Role: Computer Vision Engineer 07/2021 - 04/2022

Worked on the development of a computer vision solution applied to the industrial environment of client JBS (Seara brand), focused on automatic poultry counting in the slaughterhouse production process, using edge computing and highly complex optimization algorithms.

• Developed image processing software in C++ with the Qt framework for execution in an industrial environment with connectivity restrictions, designed to operate locally through edge computing and ensure operational efficiency without dependency on high-capacity network infrastructure.
• Implemented the Hungarian Algorithm to solve assignment problems in poultry tracking along the production line, from slaughter to evisceration, enabling automated counting and measurement of the industrial plant's daily production.
• Annotated datasets from the production line and created Deep Learning models with TensorFlow, PyTorch, and OpenCV for industrial image analysis, using MLflow for experiment tracking and CVAT for structured data annotation.
• Built model training infrastructure on Google Cloud Platform, sizing hardware for local processing in the factory and developing integrations between systems through REST APIs and PostgreSQL databases.
• Developed Proofs of Concept to validate technical hypotheses, creating new camera calibration processes through pattern projection with a projector, increasing image capture accuracy and the reliability of analyses performed by the models.

Model Training & Document Fraud Detection | GETTER S.A. / Luppa AI

Role: Computer Vision Software Engineer 02/2021 - 07/2021

Worked on the development of a document forensics platform for detecting fraud in identification documents for client Parana Banco, with contributions in machine learning, backend engineering, and system integration.

• Trained computer vision models to detect inconsistencies and signs of fraud in documents such as Brazilian ID cards (RG) and driver's licenses (CNH), preparing and organizing data for training and making iterative adjustments to improve detection accuracy and robustness.
• Developed platform integration services through REST APIs and database modeling with PostgreSQL, configuring infrastructure with a focus on operational efficiency and cost optimization.
• Participated in alignment with Parana Banco stakeholders to understand operational needs, prioritize improvements, and adjust features according to business requirements and the tool's real-world use.
• Defined and prioritized tasks for the platform's continuous evolution, contributing to the progressive improvement of model accuracy and the overall efficiency of the automated document analysis solution.

Backend Development & Data Analysis | Datatype

Role: Mid-Level Django Developer 01/2020 - 02/2021

Worked on the development of products focused on data aggregation, market analysis, and digital authentication, with a focus on solutions for the financial, insurance, and data analysis sectors.

• Developed a data aggregation product to consolidate CPF, chassis, and vehicle information, used by companies such as Suhai Seguradora and C6 Bank for risk analysis, client identification, and asset valuation processes.
• Conceived and developed an advanced electronic signature platform with legal validity, including analysis and adaptation to applicable regulatory legislation and integration with the Mailgun communication service for signature flow automation.
• Created a tool for collecting and structuring insurance market data to support analyses conducted by an IQVIA data scientist, implementing pipelines for collecting and aggregating multiple data sources.
• Defined the architecture of the developed platforms, including PostgreSQL database modeling, creation of REST APIs for integration with external systems, and development of integrations with external services.

Process Automation & Industrial Computer Vision | SoftWillians IT Solutions

Role: System Analyst 04/2019 - 01/2020

• Developed process automation robots (RPA) using Automation Anywhere combined with Python scripts for automatic data extraction from corporate systems and automation of repetitive operational workflows, reducing manual activities and increasing the reliability of processed information.
• Developed an API integrating an automatic PPE detection system with the factory incident management system, using OpenCV and YOLO for industrial image analysis and detection of non-compliance situations with automated incident registration.
• Integrated different systems in the architecture, transforming image analysis results into actionable information within the organization's operational systems and strengthening safety monitoring mechanisms in the industrial environment.

Quality Systems Engineering & Shop Floor Modernization | Ball Corporation

Role: Quality Systems Engineer 10/2018 - 04/2019

Played a key role in the deployment and evolution of the next-generation Shop Floor Quality Inspection platform, ensuring system reliability and aligning operational data with factory-wide KPIs.

• Participated in the large-scale update of the factory's quality inspection tool, supporting the transition for end users and ensuring data integrity across the new digital manufacturing ecosystem.
• Engineered and automated quality reports to track critical manufacturing targets, supporting factory-wide performance goals and strategic decision-making for plant leadership.
• Led the analysis and resolution of complex incidents within the quality management system, providing specialized technical support to quality teams across multiple factory units and reducing operational downtime.
• Managed daily operational routines and acted as a bridge between IT and the factory floor to streamline quality workflows and ensure the platform met the demands of high-volume production.

IT Service Delivery & Operations | TIVIT

Role: IT Service Delivery & Operations Specialist II 05/2016 - 04/2018

Managed first- and second-level technical support for high-scale enterprise clients, focusing on ITIL/COBIT governance, knowledge management, and the implementation of specialized remote support units for mission-critical environments.

• Engineered and led a specialized remote technical support cell to intercept complex incidents before they reached field teams, optimizing field resource allocation and reducing operational costs through remote resolution of high-complexity microcomputing issues.
• Led the end-to-end setup of service desk operations for a major mining client, establishing the knowledge base, defining standard operating procedures, and structuring incident workflows to support high first-call resolution rates.
• Provided advanced technical support for the Regional Labor Court (TRT-2), helping maintain the availability of the Electronic Judicial Process (PJe) system for judges and attorneys.
• Applied ITIL and COBIT frameworks to manage incident lifecycles, service request fulfillment, and team workload distribution while maintaining alignment with service level agreements.
• Received formal recognition from the End User Director for performance and analytical leadership during a high-pressure system crisis, with the achievement featured in the corporate monthly newsletter.

Remote Support & Service Desk Operations | TIVIT

Role: Remote Support Technician II 09/2015 - 05/2016

Managed first- and second-level technical support for enterprise clients, supporting ITIL/COBIT-based operations, knowledge management, and specialized remote support for mission-critical environments.

• Operated in a specialized remote support unit focused on resolving complex technical incidents before field escalation, improving response time and helping optimize field team allocation.
• Supported service desk operations for enterprise clients, contributing to knowledge base maintenance, incident workflow organization, and standard operating procedure execution.
• Provided technical support for institutional environments, including the Electronic Judicial Process (PJe), supporting service continuity for critical legal operations.
• Applied ITIL and COBIT practices to incident management, service request fulfillment, and workload organization in alignment with service level agreements.

Remote Support & Technical Operations | TIVIT

Role: Remote Support Technician I 09/2012 - 09/2015

Provided remote technical support for enterprise clients, working with incident management, service request fulfillment, knowledge base usage, and operational support routines in high-scale service environments.

• Handled first- and second-level technical support activities for enterprise users, resolving microcomputing incidents and escalating complex cases according to operational procedures.
• Supported service desk routines for large clients, contributing to incident workflow organization and knowledge base usage to improve resolution consistency.
• Applied ITIL and COBIT concepts to manage incidents, requests, and workload distribution in environments governed by service level agreements.
• Contributed to crisis response and operational stability during high-pressure support scenarios, supporting continuity for client environments.

Technical Skills & Knowledge

Courses & Certifications

Computer Vision with Embedded Machine Learning | Coursera

Machine Learning in Production | Coursera

Structuring Machine Learning Projects | Coursera

Time Series and Predictive Analytics: The Complete Course | Udemy

C++ Project Development | Udemy