Luis – Machine learning, Python, AWS

A global consultancy that helps the world's most ambitious change makers define the future. Luis led engineering efforts in the AI Insights and Solutions (AIS) group at the company, focusing on ML and AI applications.

• Created a event-driven system for payroll process.
• Created a RAG system for technical and scientific papers search.
• Created an ingestion process for labor contracts with keyword metadata generation.
• Spearheaded system design for AI and ML applications, ensuring seamless integration and optimal performance.

An American multinational home improvement retail corporation which sells tools, construction products, appliances, and services including fuel and transportation rentals.

• Developed 'BobAI' by fine-tuning the Home Depot text using the davinci-003 text model. Implemented FastAPI with Uvicorn for LLM interactions and utilized Airflow -> Kubeflow DSL -> Apache Beam pipeline for continuous model refinement with curated text. Employed prompt engineering techniques with a mix of blend factual and task-based queries, resulting in semantically structured responses without requiring additional handler code.
• Developed the 'Shoppable Images' architecture using Apache Beam, PyTorch, and Kubeflow pipelines.
• Designed workflow orchestration DAGs with Apache Airflow, BigQuery, and Google Cloud Storage (GCS).
• Developed Apache Beam pipelines for ETL/ELT and machine learning purposes.
• Utilized Dataflow pipelines for distributed data processing.
• Incorporated pre-trained TorchVision models for image recognition.
• Implemented cloud functions for event-driven mechanisms.

Senior data engineer for Spark .NET optimization data pipelines and build Data Lake.

• Improved Spark .NET job performance by 36% for the FA-Import project.
• Conducted proof of concept (PoC) for migrating SQL Server external DB to a data lake using S3, leveraging DMS, Glue jobs, Delta Spark, and Athena.
• Proficient in profiling and diagnosing Spark components to identify critical vulnerabilities.
• Conducted scientific research to identify critical breaking points in Spark .NET applications.

Software engineer for ML infrastructure, ML engineering and data science.

• Implemented Kubeflow pipelines architecture using Vertex AI.
• Refactored ETL pipelines replacing KFP v1 with KFP v2 (ephemeral components).
• Implemented model metadata tracker and data lineage with Kubeflow.
• Designed the 'Prism' torch model pipeline in Kubeflow and Vertex.
• Implemented services for managing data processing pipeline on Kubeflow version 1.
• Redesigned and automated a full cycle running ETL process in Kubeflow and GCP environments.
• Successfully debugged and redesigned an ETL project on Kubeflow.
• Migrated Kubeflow components from KFP v1 to KFP v2.
• Created and designed lightweight components.
• Developed event-driven functions using Cloud Functions as sensors for detecting bucket updates

Development of a new reference architecture and business applications that accelerate and enable a better customer experience, adapting the bank's products to real people.

• Optimized Spark data-mining pipeline for improved performance.
• Enhanced efficiency of Spark ML and Mllib jobs through optimization techniques.
• Engaged in Machine Learning engineering activities for distributed systems.
• Developed an Investment product type logit engine.
• Created 'M&Ms' (Minimum & Maximums) left-over balance z-score engine.
• Developed a Markov transition predictor for outlier states engine.
• Implemented Convolution filter for time series smoothing.
• Developed an engine for generating suitable trends for investment plans.
• Implemented and modeled a 360° financial health classification engine 'Check-Up'.
• Conducted research and analysis on Catalyst and Tungsten components for internal documentation.
• Managed task creation, prioritization, and automatic assignment.
• Conducted code reviews to ensure high-quality code standards.
• Collaborated with team members to discuss and review technical solutions.
• Organized local SCRUM meetings for effective project coordination.
• Worked closely with development and DevOps teams.
• Automated day-to-day operations for increased efficiency.

The company provides Flexible Human Capital, Headhunting, Payroll Processing Solutions (PPO), Business & Process Solutions (BPS), Recruitment Process Outsourcing (RPO), and Supplier Management (CWO) services.

• Led pricing and revenue models & Automated recurrent data mining pipelines with R.
• Implemented regression models for pricing prediction.
• Researched time series from historical sales factors.
• Created new sales strategies based on data analytics and hypothesis testing.
• Implemented a solver for optimizing pricing factors.
• Salesforce analytics with RForcecom library for classification models.
• Implemented training internal moocs for Salesforce adoption.