Machine Learning Engineer

LUIS
GARAY

AI & Data Science  //  Deep Learning & MLOps

Specializing in computer vision, natural language processing, and end-to-end ML systems. From fine-tuning diffusion models to deploying scalable APIs, I build AI solutions that bridge cutting-edge research and real-world impact.

9 Projects
5+ ML Domains
20+ Tools & Frameworks
Luis Garay
scroll
01 /

EXPERTISE

👁
Computer Vision
PyTorch YOLO CNN ResNet EfficientNet VGG16 Object Detection
🧠
NLP & Generative AI
BERT Transformers LangChain Stable Diffusion LoRA OpenAI HuggingFace
⚙
MLOps & Backend
MLflow FastAPI AWS EC2 scikit-learn xgboost Python
📊
Data Science
ARIMA Clustering PCA t-SNE networkx pandas matplotlib
02 /

PROJECTS

01 Bank Churn Prediction

Bank Churn Prediction

Predict customer churn in a bank using ensemble ML models, experiment tracking with MLflow, and a production API deployed on AWS EC2.

mlflow AWS EC2 scikit-learn xgboost FastAPI Python
View on GitHub
02 Langchain Virtual Assistant

Langchain Virtual Assistant

A retrieval-augmented virtual assistant powered by LangChain and OpenAI, using FAISS for semantic vector search over document collections.

LangChain OpenAI FAISS Python
View on GitHub
03 Agricultural Crop Classification

Agricultural Crop Image Classification

Benchmarking CNN, VGG16, ResNet50, EfficientNetV2, and GoogleNet for crop image classification. Deployed as a FastAPI service with MLflow tracking.

PyTorch MLflow FastAPI Python
View on GitHub
04 Sentiment Analysis

Sentiment Analysis for Amazon Reviews

API comparing BERT fine-tuning strategies for Amazon review sentiment classification, with Hugging Face Transformers and full MLflow experiment tracking.

PyTorch MLflow HuggingFace NLTK scikit-learn
View on GitHub
05
🎯

Object Detection for Stock Monitoring

Real-time object detection for shelf stock monitoring in an automated store, benchmarking YOLOv7 and NVIDIA TAO Toolkit models for optimal accuracy/speed trade-off.

YOLOv7 YOLOv7-tiny TAO YOLOv4 PyTorch
View on GitHub
06 Time Series Mauna Loa CO2

Time Series — Mauna Loa CO₂

Analysis and forecasting of Mauna Loa CO₂ atmospheric concentration over time using ARIMA, linear regression, and ML-based forecasting models.

ARIMA scikit-learn Linear Regression Python
View on GitHub
07 Gene Clustering

Clustering & Visualization for Gene Analysis

Dimensionality reduction and clustering for gene expression data, comparing PCA, MDS, t-SNE, KMeans, and Hierarchical Clustering for pattern discovery.

scikit-learn PCA t-SNE KMeans MDS
View on GitHub
08 Graph Theory Criminal Network

Graph Theory in Criminal Network

Network analysis applying graph theory to CAVIAR Investigation criminal data, identifying key nodes, centrality metrics, and network topology of criminal structures.

networkx matplotlib Python
View on GitHub
09 Generative AI Fine-Tuning

Generative AI Fine-Tuning

Stable diffusion fine-tuning on a Pokémon dataset using LoRA adapters, enabling high-quality domain-specific image generation with minimal compute overhead.

Stable Diffusion LoRA PyTorch Transformers
View on GitHub