Mario Jacobo Rios Gamboa

Enterprise GenAI architecture, multi-agent systems, and AI platform engineering. - Architected a multi-agent LLM system for credit-risk decision support using Azure AI Foundry, Microsoft Agent Framework, and Python to coordinate agents across capacity, willingness, compliance, and decision workflows. - Validated the credit-risk workflow across 50 representative scenarios, demonstrating an estimated 85% reduction in evaluation turnaround during testing. - Translated credit-decision workflows and stakeholder

Cloud AI infrastructure, backend engineering, and production GenAI services for telecom use cases. - Led the implementation of AWS infrastructure for GenAI microservices and AI product services using AWS Bedrock, ECS, S3, RDS, CloudFormation, CloudWatch, and Bitbucket CI/CD. - Improved deployment reliability by standardising reusable deployment scripts, CI/CD templates, shared Docker images, and infrastructure practices, reducing deployment time by 50%. - Built backend services and authentication workflows

AI data pipelines, transcription processing, classification systems, and large-scale telecom data automation. - Built a secure self-hosted Whisper transcription pipeline for customer-service audio, processing ~84 seconds of audio per 1 second of computation using Python, Docker, GPUs, AWS ECS, Step Functions, and batch orchestration. - Designed a RAG-based complaint classification system using Python, AWS Bedrock, LangChain, Vector Search, and LLM-based intent recognition, processing 100,000+ customer conv

Search infrastructure, data pipelines, and retail platform modernisation. - Migrated enterprise search infrastructure from AWS CloudSearch to Typesense on Kubernetes, supporting semantic and visual search across 1M+ products and reducing infrastructure costs by 83% across 37 stores. - Led the migration from StreamSets to Apache Airflow for 50+ production DAGs using Python, Helm-based deployment, GCP Cloud SQL, and BigQuery. - Improved developer productivity and reduced maintenance overhead by 40% through p

Mobile application development for retail operations. - Developed iOS mobile application features for internal retail workflows, working across application logic, API integration, debugging, and release support.

Recognized as Entel Power of the Month for June. - Achieved a 50% reduction in data ingest time to the data warehouse by implementing an efficient data pipeline using Airflow and AWS. - Applied web scraping with Selenium and BeautifulSoup to extract public market data, increasing training-data volume for existing predictive models by 20%.

Mobile application development for public-sector and enterprise field operations. - Built a Flutter mobile application for the Peruvian Ministry of Education to manage evaluation attendance, reducing logistical costs by 20%. - Developed a Xamarin.Forms mobile application for the Peruvian Ministry of Interior to support social-conflict response workflows, reducing response time by 10%. - Designed and implemented a Flutter mobile application for Suit Kote with offline registration capabilities, reducing logi

Introduction to computer science and some concepts of data structures for sophomores students

More than 1,000 Excel spreadsheets were analyzed to increase the overall experience with inside sales logistics by developing and implementing a desktop application. Including a backend endpoint and a database.

Inspected over 400 excel spreadsheets to enhance overall facility to handle sales outlet by developing a desktop application and connecting it to a sample database.

ZAB Statement of Comparability issued: degree comparable to a German university degree. Thesis: “An Empirical Investigation into Within-Project Software Defect Prediction Using CodeBERT and Static Code Metrics” Developed CodeBERT-SMCH, a hybrid Transformer architecture combining semantic embeddin

Selected for the International Baccalaureate Diploma at a High Performance College with a full academic scholarship.