Tianchang Ismael Mejia

Engineered enterprise multi-agent LangGraph and Databricks MLflow pipelines for contract analysis and NL2SQL sales analytics, reducing contract search time from hours to seconds and democratizing BI access for sales stakeholders.

Designed scalable AI pipelines (Amazon Bedrock, Claude 3.5) for invoice extraction and built video analytics and agentic AI systems to automate billing, fraud detection, and insurance workflows, improving reconciliation and enabling 24/7 automated processing.

Developed PyTorch computer vision models and MLOps infrastructure for touchless checkout, achieving 99.9% item recognition across 60K+ SKUs and reducing inference latency to 100ms while enabling weekly model updates across 700+ kiosks.

Trained and deployed perception models and MLOps pipelines for autonomous shuttles, improving detection mAP by 10% and processing 25TB/month of sensor data with automated retraining and experiment tracking.

Built ROS robotics nodes, computer vision and healthcare ML solutions, and full-stack SaaS components, improving robotic pick-and-place accuracy and deploying HIPAA-compliant ECG models with 92% AUC-ROC.

Developed recommendation models and ETL pipelines processing large customer datasets, improving CTR by 8% and enabling daily updates for product recommendation systems.

Completed a Master of Science in Computer Science focusing on advanced topics in software engineering and machine learning.

Earned a Bachelor of Science in Computer Science with coursework in algorithms, systems, and programming.