George Narita Vilarinho, Msc. Email & Phone Number

São Paulo, Sp, Br

São Carlos, São Paulo, Brasil

At Gryfo, I work as a back-end developer focused on creating new features and tests for facial biometrics solutions, maintaining and improving image processing pipelines, and refactoring code by applying best development practices to enhance application performance and reduce cloud computing costs (AWS).Key Achievements: - Migrated the image embedding creation service from EC2 to a serverless solution (AWS Lambda), resulting in significant cost reduction and increased service scalability.- Achieved nearly 3 times faster server-client record synchronization after refactoring the code handling DynamoDB calls.- Enabled near real-time detection of API call failures through the implementation of a constantly running health-check service.Main Activities:- Developed and optimized serverless APIs for highly scalable and available AI models using AWS Lambda and API Gateway.- Created and improved AI pipelines for real-time image processing, increasing application efficiency and reducing response time.- Implemented health-check routines and serverless solution monitoring to ensure service continuity.- Applied clean code and refactoring practices to promote sustainable and maintainable code.- Implemented unit and performance tests with pytest to validate application scalability before deployment.- Used AWS (S3, SQS, SNS, EC2, ELB, DynamoDB, Global Accelerator, CloudWatch) for infrastructure and monitoring.- Collaborated in development using Git and GitLab.- Experienced in agile methodologies: Scrum, Kanban, and XP.Technologies: Python, Flask, Requests, Git, Pandas, Numpy, OpenCV, Pillow, pytest, AWS.

São Paulo, São Paulo, Brasil

At Verx, I worked on a project at Serasa Experian for nearly 3 years. I was responsible for implementing and maintaining web scrapers for Serasa Experian's main data acquisition application, as well as the entire public data ETL pipeline. I also supported Data Science projects focused on NLP and computer vision. Additionally, I was part of the Agro data squad, dealing primarily with geospatial data and operations.Key Achievements:- Implemented an hCaptcha solver, enabling automated data capture from sources requiring Captcha resolution.- Actively contributed to migrating all web crawler and scraper code from JavaScript to Python for easier maintenance by the predominantly Python developer team.- The migration also facilitated ETL workflow automation within the squad.- Automated the data enrichment process and generated reports and studies, adding value to the business unit and fostering future project engagements.Main Activities:- Contributed to the development and maintenance of ETL pipelines for data ingestion and processing, ensuring efficiency and automation.- Developed web scrapers using Selenium, Pypeteer, and Playwright for automated data capture.- Manipulated geospatial data with GeoPandas and PostGIS, performing geometry manipulation, spatial queries, and aggregations.- Utilized AWS services like Airflow, S3, RDS, Lambda.- Developed and optimized APIs using Flask and FastAPI.- Documented APIs with Swagger.- Maintained web crawlers implemented in Node.js.- Fine-tuned pre-trained computer vision models to efficiently generate embeddings.- Engaged in all project lifecycle phases, from image collection and data labeling to model benchmarking and classifier performance optimization.- Deployed models as cloud-based microservices.Technologies: Python, Git, Pandas, GeoPandas, Dask, Selenium, Playwright, Pyppeteer, requests, Flask, FastAPI, PostgreSQL, PostGIS, Airflow, AWS, Tensorflow, OpenCV, Regex e Pillow.

São Paulo E Região, Brasil

As a data scientist at NTT DATA, I worked on Natural Language Processing (NLP) projects for multilingual chatbots. Key Achievements:- Implemented the first version of the virtual assistant's entity recognizer, a highly requested feature by the company's linguists.- Replaced the chatbot's cognitive engine by moving from Word2Vec pre-trained models to a single multilingual LLM (BERT), significantly improving intent classification accuracy and reducing memory allocation costs on GCP instances.- Reduced chatbot response time by applying model distillation and switching the backend from TensorFlow to PyTorch.Main Activities:- Developed NLP solutions for multilingual chatbots, working on tasks such as intent classification, entity recognition, question and answer (Q&A), and topic recognition to improve user interaction.- Created and maintained data pipelines on Google Cloud Platform (GCP) with microservices architecture, ensuring efficient data flow between services.- Conducted exploratory analysis and applied data balancing techniques.- Labeled datasets with Doccano and SpaCy.- Leveraged transfer learning and fine-tuning to adapt models to new domains and improve prediction accuracy.- Conducted grid search for hyperparameter optimization, enhancing machine learning model performance.- Applied model distillation techniques to reduce runtime resource needs without significantly compromising performance.- Converted trained models to ONNX format for compatibility with different frameworks and performance optimization in production environments.Technologies: Python, Regex, Pandas, Git, Pandas Profiling, Seaborn, Matplotlib, Plotly, Dask, Doccano, Scikit-learn, Tensorflow, Keras, pyTorch, SpaCy, Polyglot, AllenNLP, regex, NLTK, Gensim, DeepPavlov, XGBoost, SBERT, Hugging Face, Docker, Kubernets, Google Cloud Platform, MongoDB.

São Carlos E Região, Brasil

Brazilian AI startup which uses Computer Vision to detect and recognise people and objects through cameras. Gryfo offers a complete and customisable platform, which makes processes much safer and more reliable in real use situations, like: tracking people, vehicles and others objects; emotion recognition, people counting, heatmap, smart alerts and etc. Active participation in projects involving:- Machine Learning- Computer Vision- Information Retrieval- Natural Language Processing- Text ClassificationResponsabilities:- Planning and development of projects involving Computer Vision and NLP following GitFlow;- Implementation of software tests;- eXtreme Programming;- Activity management using YouTrack;- Implementation and homologation of systems;- Presentation and demonstration of products in events;- OKR-based management methodology;- Experience in SCRUM;

Ribeirão Preto E Região, Brasil

Supervised internship conducted with the subject of Algorithms and Data Structure I for the Biomedical Informatics course with workload of 6 hours/week. The emphasis of the course was to introduce students to the concepts and operation of the most important data structures implemented in C / C ++. Responsabilities:- Classes attendance;- Teaching classes supervised by the subject teacher; - Assistance to students at special times to answer questions;- Homework and exercise proposals;- Assisting the teacher in preparing tests and exercise lists; - Exercise correction;

Ribeirão Preto E Região, Brasil

Supervised internship conducted with the subject of Introduction to Object Oriented Programming for the course of Biomedical Informatics with workload of 6 hours/week. The emphasis of the course was to introduce students to the concepts of the object-oriented programming paradigm with Java.Responsabilities:- Classes attendance;- Teaching classes supervised by the subject teacher; - Assistance to students at special times to answer questions;- Homework and exercise proposals;- Assisting the teacher in preparing tests and exercise lists; - Exercise correction;

Recife E Região, Brasil

Motorola is one of the world’s fastest growing smartphone providers, creating affordable, high-quality products and enabling mobile commerce in emerging markets.Responsibilities:Worked with a large variety of mobile applications and led the test life cycle of one of those applications. Provisioned daily quality reports, released Play Store upgrades as well;- Execute manual tests for Motorola mobile applications such as functional, usability, KPI, regression, live test, unity and others;- Creation and maintenance of test plans and test cases;- Log analysis of the applications;- Creation and maintenance of testing documents;- Defect tracking, using test software;- Worked closely with Test Automation specialist in identifying and creating test scenarios/test cases for automation;- Meetings with other test leads to come out with innovative ideas regarding testing methodologies;

Recife E Região, Brasil

Motorola is one of the world’s fastest growing smartphone providers, creating affordable, high-quality products and enabling mobile commerce in emerging markets.Responsibilities:- Execute manual tests for Motorola Moto Voice application such as sanity, KPI, regression, unity;- Defect tracking using test software;

Teresina

Rich Publicidade is an advertising agency that uses technology to leverage business to achieve results with imagination, creativity and market analysis.Responsabilities:- Websites development using WordPress;Front-End skills: HTML5, CSS3, JavaScript, JQuery and PHP;

Universidade Federal Do Piauí

Responsabilities:- IT Support Help Desk (12 months)- Web development using PHP (Zend 2), HTML, CSS, JavaScript (12 months)

Msc Student, Applied Computing

Software Residency In Test Analysis - Motorola, Software Engineering - Quality Assurance

Bachelor’S In Computer Science, Information Technology