Full Stack AI Developer | Intelligent Document Processing

Sherbrooke, Québec, Canada

yagmur.gulec@outlook.com · Portfolio · GitHub · LinkedIn

Full-stack AI developer with experience building and deploying intelligent document processing systems in a non-profit environment. Specialized in machine vision and OCR-based table extraction from PDF files, with a strong focus on AI quality assurance through hallucination measurement, temperature tuning, and experiment tracking. Comfortable working across the entire development lifecycle from rapid prototyping to production deployment.

• Frontend: React
• Backend: FastAPI (Python)
• AI/ML: OCR (Optical Character Recognition) for PDF table extraction, Large Language Models, hallucination detection, temperature parameter optimization
• LLM Observability: hallucination detection, temperature tuning, prompt engineering, MLflow experiment tracking, OpenSearch validation
• Deployment: Railway
• Other: REST APIs, PDF processing, table extraction pipelines

Programmer-analyst — Observatoire Global du Saint-Laurent Oct 2025 – Present

• Develop full-stack AI applications using FastAPI (backend) and React (frontend)
• Achieved 90%+ hallucination-free LLM outputs in production through temperature tuning and observability pipelines
• Measure and mitigate AI hallucinations using text search and temperature parameter adjustment
• Track and debug AI model quality using MLflow (experiment tracking, performance monitoring)
• Deploy applications and rapid prototypes using Railway
• Integrate AI models into complete, user-facing production systems via REST APIs

Machine Learning Developer Intern — M2M, Remote May 2025 – Jun 2025

• Enhanced YOLO-based wildfire detection with domain-specific data and augmentation pipelines
• Applied computer vision techniques for image preprocessing and labeling automation

Software Developer — Riipen Beyond the Cloud and Level UP, Remote Feb 2024 – Mar 2025

• Developed a dashboard for prediction market data and deployed for the customer's (AWS Lambda, SAM, Python, Dash)
• Automated provisioning Jenkins Server with Terraform using Nginx on AWS EC2
• Feedback and Projects — Riipen Profile

Research Assistant in Mechanical Engineering — Université de Sherbrooke, Québec Feb 2019 – Sep 2021

• Developed numerical models to simulate a vapor bubble in a boiling liquid pool (Object-oriented C++)
• Deployed and automated large-scale simulations on HPC cluster on Calcul Québec (Bash, Linux)
• Implemented a Java plugin to process images to measure bubble sizes (Java, ImageJ)

Master's in Computer Science — Bishop's University, Sherbrooke, QC Jan 2022 – May 2024 GPA: 3.9/4.0 | Courses: Data Mining, Database Design, Deep Learning