Machine Learning Engineer

ETRON HATTON

Machine Learning Engineer

Summary

Dynamic and results-driven AI/ML professional with over 11 years of industry experience designing, developing, and deploying advanced machine learning solutions across NLP, LLMs, and Computer Vision. An expert in building end-to-end ML pipelines, including chatbot systems, document intelligence solutions, and visual recognition applications. Adept at applying deep learning architectures (Transformers, CNNs, LSTMs, GNNs), statistical models, and large language models (GPT, LLaMA) to solve complex real-world problems.

Equally skilled in engineering best practices, leveraging modern cloud platforms (AWS, GCP, Databricks), MLOps tools (SageMaker, MLflow, Docker, Kubernetes), and cutting-edge frameworks like Hugging Face, LangChain, and PyTorch. Strong background in data engineering with experience in Snowflake, Kafka, and dbt.

Education

University of North Texas

Master's Degree in Computer Science 2013 - 2014

Texas State University

Bachelor's Degree in Computer Science 2008 - 2012

Experience

AWS Seattle, WA

Machine Learning Engineer 08/2021 - Present

Led the development and optimization of Amazon Elastic Inference, integrating machine learning frameworks like TensorFlow, Apache MXNet, and PyTorch. Restructured large language models (LLMs), improving inference speed by 57% and reducing memory usage by 30%, resulting in cost savings of 61% through optimized resource utilization. Led the development of Amazon Kendra, integrating NLP and DL models like BERT and T5 for intelligent search. Streamlined integration with AWS services (SageMaker, OpenSearch, Polly) to deliver scalable, efficient search solutions. Optimized for performance and cost-efficiency, reducing query response times and infrastructure overhead, while boosting user satisfaction. Contributed to a 25% increase in AWS’s AI/ML services revenue.

Architected scalable ML pipelines using LLMs (LLaMA), LangChain, and Hugging Face Transformers for intelligent code analysis, semantic search, and NLG. Optimized performance, reducing processing time by 40% and cutting cloud infrastructure costs by 35%. Automated key processes, reducing manual review time by 50% and boosting team productivity. These improvements led to faster deployment cycles, a 30% increase in throughput, and significant cost savings across the pipeline. Meta Menlo Park, CA

AI/ML Engineer 07/2017 - 07/2021

Led the development of BlenderBot using NLP and Generative AI models, fine-tuning BERT for tasks like text classification, sentiment analysis, and entity recognition. Reduced training time by 35% using distributed training and early stopping, cutting compute costs by 40% while maintaining or improving accuracy. Streamlined model development, achieving 20% faster time-to-market for production-ready models, driving significant cost savings in both development and operations, and increasing economic efficiency for scalable, cost-effective NLP solutions. Implemented RAG pipelines for enhanced information retrieval in LLM-based systems, improving response accuracy and context-awareness. Reduced response time by 30% through efficient indexing and caching, cutting compute costs by 25% while maintaining or improving response quality. Automated retrieval, boosting throughput by 40% and contributing to operational efficiency and significant cost savings in deployment and maintenance.

Integrated Apache Kafka for real-time data streaming and pipeline orchestration, enhancing model inference and monitoring. Reduced latency by 35% and improved processing efficiency, cutting infrastructure costs by 40% during peak loads. Automated the data pipeline, increasing throughput by 55%, driving operational efficiency and significant cost savings in data processing and monitoring. www.enhancv.com Powered by

•

•

•

•

•

•

************@*****.*** linkedin.com/in/etron-hatton-227954362 Houston, TX 77016 Experience

ProData Infotech Atlanta, GA

AI Developer / Intern 11/2014 - 07/2017

Developed and optimized computer vision models for object detection and image classification using CNN architectures like YOLO and VGG. Improved accuracy and reduced computational costs through techniques like model pruning, quantization, and mixed-precision inference. Reduced inference time by 40% and resource usage by 35%, enhancing scalability and lowering infrastructure overhead. These optimizations led to a 25% reduction in operational costs while maintaining performance, driving cost savings and improving economic efficiency. Preprocessed large-scale image datasets using OpenCV and Python for tasks such as face recognition and image segmentation. By streamlining the data pipeline, the overall data preparation time for machine learning models was reduced by 25%, contributing to faster model training cycles and enhanced operational efficiency.

Designed and implemented OCR pipelines for document parsing using Tesseract and custom deep learning models, enabling automated data extraction from scanned forms, increasing the accuracy as 98%. Skills

Machine Learning & AI Models: GPT LLaMA CNN RNN GNN RAG Tools & Frameworks: Hugging Face PyTorch TensorFlow MLflow Docker Kubernetes Snowflake Kafka Programming Languages: Python C# PHP Matlab SQL HTML JavaScript Frameworks & Libraries: LangChain LangGraph Pycharm Django FastAPI Flask Achievements

Amazon Elastic Inference

Integrated machine learning frameworks and

restructured LLMs and developed and

deployed mixed-precision techniques and

model quantization, resulting in achieving

scalable, high-performance solutions.

Amazon Kendra

Led the integration of NLP and DL models

and orchestrated the seamless integration

with AWS services resulting in 30% increase

in AWS's revenue within the AI/ML services

sector.

BlenderBot

Led the development of ML models focused

on NLP and GenAI, including fine-tuning the

transformer model BERT, resulting in a 40%

reduction in compute costs and a 20% faster

time-to-market for production-ready models.

Certification

Python for Data Science and Machine Learning Bootcamp, Udemy, 2021 Programming Essentials in Python, Python Institute, 2020 www.enhancv.com Powered by

•

•

•

Contact this candidate