Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

Computer Vision & MLOps Engineer with production experience designing, deploying, and optimizing LLM and RAG systems (Azure-hosted Llama 3 + vector stores) alongside GPU inference workloads on NVIDIA T4. Built end-to-end pipelines for data ingestion, model serving, evaluation, benchmarking, monitoring, and automated retraining using Kafka, Airflow, Terraform, and auto-scaling GKE. Partnered cross-functionally with Data Engineering, Product, and Analytics teams to deliver measurable business impact through production AI systems—enforcing sub-100ms P95 latency while cutting third-party costs by 60% and reducing release cycles to under 15 minutes.…Computer Vision & MLOps Engineer with production experience designing, deploying, and optimizing LLM and RAG systems (Azure-hosted Llama 3 + vector stores) alongside GPU inference workloads on NVIDIA T4. Built end-to-end pipelines for data ingestion, model serving, evaluation, benchmarking, monitoring, and automated retraining using Kafka, Airflow, Terraform, and auto-scaling GKE. Partnered cross-functionally with Data Engineering, Product, and Analytics teams to deliver measurable business impact through production AI systems—enforcing sub-100ms P95 latency while cutting third-party costs by 60% and reducing release cycles to under 15 minutes.

Joshue Rogelio Garcia Butron

AI Engineer, Data Scientist, Data Engineer, +2





Computer Vision & MLOps Engineer with production experience designing, deploying, and optimizing LLM and RAG systems (Azure-hosted Llama 3 + vector stores) alongside GPU inference workloads on NVIDIA T4. Built end-to-end pipelines for data ingestion, model serving, evaluation, benchmarking, monitoring, and automated retraining using Kafka, Airflow, Terraform, and auto-scaling GKE. Partnered cross-functionally with Data Engineering, Product, and Analytics teams to deliver measurable business impact through production AI systems—enforcing sub-100ms P95 latency while cutting third-party costs by 60% and reducing release cycles to under 15 minutes.…Computer Vision & MLOps Engineer with production experience designing, deploying, and optimizing LLM and RAG systems (Azure-hosted Llama 3 + vector stores) alongside GPU inference workloads on NVIDIA T4. Built end-to-end pipelines for data ingestion, model serving, evaluation, benchmarking, monitoring, and automated retraining using Kafka, Airflow, Terraform, and auto-scaling GKE. Partnered cross-functionally with Data Engineering, Product, and Analytics teams to deliver measurable business impact through production AI systems—enforcing sub-100ms P95 latency while cutting third-party costs by 60% and reducing release cycles to under 15 minutes.

Available to hire

Computer Vision & MLOps Engineer with production experience designing, deploying, and optimizing LLM and RAG systems (Azure-hosted Llama 3 + vector stores) alongside GPU inference workloads on NVIDIA T4. Built end-to-end pipelines for data ingestion, model serving, evaluation, benchmarking, monitoring, and automated retraining using Kafka, Airflow, Terraform, and auto-scaling GKE.

Partnered cross-functionally with Data Engineering, Product, and Analytics teams to deliver measurable business impact through production AI systems—enforcing sub-100ms P95 latency while cutting third-party costs by 60% and reducing release cycles to under 15 minutes.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Intermediate

Intermediate

Intermediate

Beginner

Beginner

Beginner

Beginner

Beginner

Beginner

Language

Spanish; Castilian

Fluent

English

Fluent

Work Experience

MLOps Engineer at InTheLoop

July 1, 2025 - December 1, 2025

Deployed production multi-label classification and background-removal systems for retail attribute prediction on auto-scaling GPU-backed GKE, outputting structured JSON for downstream pricing and listing workflows. Fine-tuned vision-language models (SFT) on apparel tag imagery to extract entities (brand, size, material) for automated inventory enrichment and pricing intelligence. Collected and monitored GPU telemetry for NVIDIA T4 GPUs into Prometheus and Grafana. Engineered Kafka and Airflow streaming pipelines for real-time image ingestion, GPU inference, model training, and automated retraining. Provisioned production GCP environment with Terraform (private GPU-backed GKE, VPC isolation, IAM, load balancing, Artifact Registry) and built CI/CD pipelines (GitHub Actions + Docker + K6) to reduce release cycles to under 15 minutes with automated rollback. Designed and deployed a fault-tolerant three-tier caching architecture that reduced third-party API costs by 60% while enforcing su

Experience Machine Learning Engineer at NTT Data

November 1, 2023 - Present

Designed and built a production RAG system over large U.S. federal regulatory corpora using custom scrapers, Azure-hosted Llama 3, Cosmos DB vector store and Blob storage. Measured and minimized hallucinations while evaluating context relevance, response fidelity, answer relevance, and faithfulness; configured Azure Monitor & Application Insights for continuous monitoring. Delivered source-grounded answers with direct verification URLs, reducing research time by at least 1 hour per query and enabling chamber agents to handle more compliance tickets. Fine-tuned Tesseract OCR and engineered a scalable extraction pipeline to process highly variable audit PDFs, including clean digital to low-quality photocopies. Implemented keyword/context-based parsing with deterministic validation and exception isolation, reducing CER and WER by 95%, cutting manual review time by 80%, and increasing compliance throughput. Collaborated with Data Engineering to design and optimize terabyte-scale data ext

Machine Learning Engineer at FRINNERT

May 1, 2022 - November 1, 2023

Designed a tax-automation pipeline combining OCR with anomaly detection models (Isolation Forest, KNN) to process high volumes of XML invoices and identify fraudulent patterns in CFDI billing flows. Mentored and led a team of 5 junior/recent-graduate engineers to deliver the complete annual tax-declaration workflow, establishing code review practices, testing standards, and documentation while ensuring end-to-end module quality from data validation to reporting.