AI Engineer | Machine Learning Engineer | Generative AI Specialist Based in 📍 Toronto, ON
🔗 linkedin.com/in/javier-medel
AI Engineer with 5+ years of experience delivering production-ready solutions in computer vision, IoT, and Generative AI. Skilled in LLMs, vector databases, and CI/CD pipelines to build scalable, data-driven applications. Passionate about creating practical AI systems that enhance productivity and drive measurable business impact.
AI Engineer / Senior Machine Learning Engineer | Nov 2022 – Present
- Breakdown Interface: Main contributor to an AI-powered system integrating IoT data with CRM systems to support trailer order creation and material classification. Applied statistical methods to detect outliers in bale weights, streamlining the order process.
- Bale Segmentation System: Developed object detection (YOLOv11, Python) and segmentation workflows using Vision-Language Models (SAM), improving accuracy and operational efficiency.
- RAG Knowledge System: Built a Retrieval-Augmented Generation solution using LLMs, LangChain, and vector databases (FAISS, Pinecone) to allow operational staff to query manuals and procedures in real time.
- Model Deployment & Monitoring: Deployed models on Azure using APIs, adding observability with production metrics tracking.
- Data Analytics & Reporting: Leveraged SQL for ETL, analytics, and insights; created decision-driving dashboards using Power BI.
Tech: Python, PyTorch, YOLO, VLM, GPTs, LLMs, Vector Databases, SQL, Azure ML, Power BI, LangChain
Lead Data Scientist / Sr Machine Learning Engineer / Machine Learning Engineer | Feb 2019 – Nov 2022
- Visual Monitor: Designed and deployed YOLOv7 object detection models across 300+ Jetson Nano IoT devices, containerized with Docker for scalable edge deployment.
- Baler Validation System: Built ensemble classification workflows (Random Forest + image models) to validate recyclable bales, reducing manual validation and improving accuracy.
- Implemented outlier detection to flag low-weight or misclassified bales, reducing manual checks by 35%.
- Fullness Monitor: Developed forecasting models using Python/Keras to optimize container collection, cutting costs by 12%. Integrated insights into Power BI.
- Applied robust feature engineering to boost model performance across supervised and unsupervised workflows.
Tech: Python, TensorFlow, YOLO, Power BI, Azure, Clustering, Docker, Statistics, Machine Learning, Scikit-Learn
Deep Learning: NLP Researcher | Jul 2018 – Dec 2018
Data Management Analyst (Retail & Ecommerce) | May 2016 – Sep 2017
Data Analyst ETL (Banking Domain) | Feb 2015 – May 2016
- Built an AI-powered practice platform for the Canadian citizenship test using LLMs and Generative AI.
- Orchestrated AI workflows with Cursor AI and deployed the web app on Vercel.
- Used OCR and web scraping to digitize the official study guide for accurate question coverage.
- Implemented vector database pipelines for context-aware question retrieval.
- Set up CI/CD workflows for automated testing and deployment.
Tech: GPTs, Vector Databases, Generative AI, Cursor AI, Vercel, Python, OCR, Web Scraping, CI/CD, SQL
- MSc Data Science, St George’s, University of London (2018)
- AI Specialization, University of Calgary (2022)
- Academy Accreditation – Generative AI Fundamentals, Databricks (2025) — ID 163715660
- AI Engineer using Microsoft Azure Nanodegree, Udacity (2022)
- Industrial IoT: From Theory to Applications, MIT (2022)
Computer Vision, Neural Networks, Predictive Modeling, Feature Engineering, Deep Learning
GPTs, LLMs, Vector Databases, Retrieval-Augmented Generation (RAG)
SQL, Azure ML, Power BI, IoT Data Integration
PyTorch, YOLOv7/v11, SAM (VLM), Cursor AI, Vercel, CI/CD Pipelines
Cross-functional teamwork, Stakeholder Engagement
- English: Professional Proficiency
- Spanish: Native