minh son

At Tomota Otanics / Otanics Technology JSC, I tuned multi-tier SQL queries on IoT sensor and pond-log records, cutting query runtime from over 45 seconds to 4.2 seconds. I also built Python ingestion pipelines to sync Redash REST APIs into PostgreSQL 16.

I established Data QA/QC protocols and a Data Completeness Score for historical crop-cycle data, and created a standardized Data Dictionary after auditing the production workbook. For shrimp transfers, I modeled lineage across ponds and found a correlation between high Delta Alkalinity and post-transfer mortality spikes; the finding was integrated into corporate SOP transfer guidelines.

I applied Panel Data Fixed Effects and Kruskal-Wallis tests to analyze probiotic dosage, bacterial counts, and mortality, confirming Probiotic B2W4's disease-suppressing efficacy and optimal dosage. I also analyzed equipment telemetry to identify how unmonitored over-feeding preceded toxic nitrogen peaks and added operational costs.

For my Tomota Farm Copilot capstone, I’m developing an AI multi-agent decision-support platform that connects IoT telemetry with SOP-grounded Agentic RAG and deterministic biochemical calculations. My other projects include a D3.js student analytics dashboard, LSTM stock forecasting and anomaly detection, and a resume-job matching platform.