Arnold Madayag

Built backend services in Python and FastAPI using PydanticAI for structured agent development and reliable tool execution. Developed an AI orchestrator that routes requests to specialised agents for hiring, compliance, fleet operations, document processing, driver onboarding and task management.

Built an MCP server over the client's public API, generating tool schemas from the OpenAPI spec and grouping more than a hundred tools by domain with tool search. Built a multi agent dispatch workflow with planner, checker, and human approval, and an analytics assistant over BigQuery with semantic layer and dry run validation.

Designed and built AI modules for environmental prediction and anomaly detection with scikit learn and TensorFlow. Built models for irrigation scheduling, climate forecasting and crop health analysis, and a question answering assistant over farm data and agronomy guides with RAG.

Designed a machine learning algorithm that clustered users by geolocation for proximity based grouping. Built GPS tracking and courier assignment features for deliveries, and a pipeline that transcribed customer calls with Cloud Speech to Text and ran sentiment analysis.

Built real time analytics features for ad performance dashboards. Processed large datasets in real time with AWS Lambda and DynamoDB, and built REST APIs that pulled and combined data from several sources.

Built medical computer vision and machine learning solutions, including CNN based inspection systems. Designed signal processing pipelines that turned raw sensor data into clean time series, and integrated patient monitoring data into a live dashboard with alerts.

Supported feature development, bug fixing and testing for enterprise financial software products. Wrote SQL queries and reports, fixed defects in data import modules, and documented features and fixes for the support team.

Pursued a Bachelor's Degree in Computer Science from 2010 to 2014.