Rafael Dolores is a Software Engineering student at York University - Lassonde School of Engineering with experience in research development, software infrastructure, data engineering, and software engineering internship…
Rafael Dolores is a Software Engineering student at York University - Lassonde School of Engineering with experience in research development, software infrastructure, data engineering, and software engineering internships. He has worked with Python, Django, AWS, Docker, Kubernetes, React, Flask, and machine learning tools to build data pipelines, web applications, and automation systems.
Assisted in the research and development of Db2une, a system for automatic configuration parameter tuning for IBM DB2 to maximize query workload performance using Deep Reinforcement Learning.
Contributed to building a training pipeline for Db2une using PyTorch, Gensim, and Scikit-Learn.
Built a corresponding Django web application for training visualization, parameter inference, and model performance comparisons.
Developed a data processing infrastructure to scrape and contextualize federal court data using Beautiful Soup, Selenium, Apache Flink, Pandas, and Dask to support in-depth analysis by lawyer stakeholders.
Orchestrated containerized application workflows with Kubernetes and Docker, paired with Prefect for task scheduling, and leveraged GitHub Actions to automate monthly redeployment of cluster resources.
Provisioned and managed cloud resources on OpenStack platform using Terraform to improve deployment times and reduce manual configuration errors.
Automated legal case categorization by training a Longformer model using PyTorch and Hugging face Transformer, integrated with Langchain and OpenAI APIs.
Achieved a 50% reduction in manual lawyer review workload while maintaining OpenAI token usage within budget constraints.
Developed event driven infrastructure using Serverless Framework combined with AWS Lambda, Amazon EventBridge, and CloudFormation to transmit alarm signals from communicator panels to a central station via cellular network.
Deployed backend services using Node.js, Amazon API Gateway, DynamoDB, and AWS Secrets Manager to facilitate dealer account activation and deactivation processes and firmware updates on dealer devices.
Optimized cross-platform mobile applications performance by implementing efficient algorithms and leveraging platform-specific optimizations in React Native and Flutter.
Resulted in a 40% reduction in app load time.
Upgraded legacy applications to transition users accounts from .NET memberships to Okta using ASP.NET MVC and C# for codebase enhancements.
Architected an ETL data pipeline to automate laboratory data migration from Sanofi’s MSSQL databases into AWS S3 buckets.
Reduced 150 hours of manual production work per year using pandas library in Python.
Redesigned a decade-old MATLAB program that categorizes raw sensor data based on vaccine metrics.
Developed an equivalent full-stack web application using Flask, React, and PostgreSQL.
Reduced yearly expenses by $20000 through the removal of the MATLAB license.
Improved the application’s performance by implementing a multi-class classification model with Random Forest Classifier using Scikit-learn library.
Increased categorization accuracy from 77% to 96%.
Supported data modeling team in maintaining STEM, a Django application for monitoring supply chain data, by writing new views, securing APIs with JSON Web Tokens, and implementing user authentication.
نظام التوظيف التشغيلي المعتمد على الذكاء الاصطناعي أولاً، الذي يوحّد ويؤتمت دورة التوظيف بالكامل للمؤسسات الحديثة حول العالم.
د.جوب منطقة حرة ذ.م.م. 2026 © جميع الحقوق محفوظة
