Data Scientist | Machine Learning Engineer

📍 Lipa City, Philippines
📧 kyellpublico@gmail.com
📱 +63 9062002179
🔗 GitHub: https://github.com/Kyellpublico
🔗 LinkedIn: https://www.linkedin.com/in/edsequille-publico

Data Scientist and Machine Learning Engineer with hands-on experience in developing predictive models, early warning systems, and data-driven decision support tools using open-source technologies. My work focuses on public health, disaster risk reduction, and applied AI research, particularly in climate-sensitive forecasting and public-sector applications. Experienced in end-to-end machine learning workflows including data preprocessing, model development, evaluation, documentation, and deployment-ready prototyping.

• Python
• C/C++
• Dart

• Predictive Modeling
• Classification
• Time-Series Analysis
• Anomaly Detection
• Feature Engineering
• Model Evaluation and Calibration

• LSTM
• XGBoost
• YOLO (Object Detection)
• Vision-Language Models (VLM)
• Retrieval-Augmented Generation (RAG)

• PyTorch
• TensorFlow
• Scikit-learn
• Pandas, NumPy, Matplotlib
• Hugging Face
• Kaggle
• Google Colab
• Vector Databases
• Roboflow

• Climate and Environmental Data
• Public Health Analytics
• Disaster Risk Modeling
• GIS-Integrated Datasets

Focus: Public Health | Predictive Modeling

Developed a machine learning-based dengue outbreak forecasting system using climate and environmental indicators. The system provides 1–8 week early warnings to support public health planning and risk assessment.

Key Contributions:

• Processed and analyzed climate-sensitive health datasets (Project C-CHAIN)
• Built and evaluated predictive models using LSTM and XGBoost
• Engineered lagged and rolling climate features for time-series forecasting
• Produced model evaluation metrics and documented results for interpretability
• Designed the system as a decision-support tool for health surveillance

🔗 GitHub: https://github.com/Kyellpublico/dengue-defensor

Focus: Public Safety | Disaster Risk Reduction

Designed a location-aware disaster prediction system using stochastic modeling and satellite-based weather data to support early warning and preparedness efforts.

Key Contributions:

• Integrated satellite weather and environmental datasets
• Implemented stochastic XGBoost models for flood risk prediction
• Focused on realistic model behavior through probabilistic approaches
• Structured outputs for practical early warning and planning use

🔗 GitHub: https://github.com/Kyellpublico/ligtas-ai

Focus: HR Technology | NLP & LLMs

Built a Retrieval-Augmented Generation (RAG) system for semantic resume parsing and automated job-fit gap analysis.

🔗 GitHub: https://github.com/Kyellpublico/ai-resume-matcher

Focus: Healthcare Automation | Natural Language Processing

Developed an NLP-based system using Large Language Models to automate medical documentation workflows and transcription organization.

🔗 GitHub: https://github.com/Kyellpublico/Healthcare-NLP-Automation-using-LLMs

Researched and implemented fine-tuning and inference optimization strategies for Qwen VL 2.5 (72B), reducing inference latency from approximately 17 seconds to under 10 seconds using 4-bit quantization and QLoRA. Focused on deployment efficiency and performance constraints.

Collected and annotated datasets using Roboflow, applied data augmentation techniques, and trained a YOLO model achieving 84% mAP for food ingredient detection tasks.

• AI Engineer for Developers Associate – DataCamp
• Data Scientist Associate – DataCamp

🔗 Certificate Links:

• https://www.datacamp.com/certificate/AIEDA0015849367842
• https://www.datacamp.com/certificate/DSA0017627754099

• Cross-platform mobile development using Flutter and Dart
• Deployment-focused AI system design

Building AI-driven tools to support public health, disaster resilience, and data-informed decision-making for local communities.