Marthan Lanuzga

AI/ML & Full-Stack Developer
Caloocan, PH.

About

Highly analytical and results-driven Computer Engineering student with a strong foundation in AI/ML, computer vision, and full-stack development. Proven ability to design and optimize complex data pipelines, architect AI-powered solutions, and develop robust, scalable applications. Seeking to leverage expertise in machine learning, system architecture, and full-stack development to drive innovation and deliver high-impact technical solutions in a dynamic environment.

Work

Sasquatch
|

Lead Developer

Metro Manila, Metro Manila, Philippines

Summary

Spearheaded the development of a behavior-aware AI coding assistant, focusing on real-time feedback and scalable architecture.

Highlights

Built a behavior-aware AI coding assistant as a VS Code extension, monitoring live coding behavior to provide deterministic inline feedback for syntax, type, and runtime errors.

Engineered local deterministic analysis for immediate feedback on common coding errors, reducing latency below 50ms for enhanced developer experience.

Designed a deterministic-first architecture with clear separation between local analysis and a future AI escalation layer, ensuring scalability and modularity.

Med-ID
|

Backend Developer

Summary

Architected and engineered a hybrid computer vision pipeline for drug label analysis, ensuring high accuracy and data compliance.

Highlights

Architected a hybrid computer vision pipeline integrating PaddleOCR with a custom-trained spaCy NER model, achieving 96% text extraction accuracy on high-noise, low-light medical imagery.

Engineered a verification layer that cross-referenced OCR outputs against the OpenFDA database, ensuring 100% data compliance with safety standards.

Integrated Google Gemini 2.5 API for summarization and semantic analysis, converting raw technical drug data into patient-friendly explanations.

DOST - Advance Science and Technology Institute
|

Software Development Intern

Quezon City, Metro Manila, Philippines

Summary

Led the design and optimization of a modular data processing pipeline for computer vision, improving efficiency and ensuring compliance with government standards.

Highlights

Designed a modular data processing pipeline in Python, reducing manual feature extraction time by 40% for a large-scale computer vision project.

Optimized ML pipelines by replacing arbitrary parameters with ISO-compliant thresholds, ensuring 100% alignment with validated government standards.

Collaborated within an Agile team of 9, integrating React interfaces with ML models and delivering key features 10% ahead of schedule.

Education

Catanduanes State University
Catanduanes, Catanduanes, Philippines

BS

Computer Engineering

Skills

Programming Languages

Python, SQL, JavaScript, HTML/CSS, C++, TypeScript.

AI/Machine Learning

TensorFlow, PyTorch, Keras, OpenCV, pandas, RAG Systems, AI Agent Orchestration, Autonomous Systems, Computer Vision, Natural Language Processing, Semantic Analysis, NER Models.

Web Development

React, Vue, FastAPI, React.js, Vue.js, Vite, Full-Stack Development.

Databases

PostgreSQL, ChromaDB, SQLite, MySQL.

Tools & Platforms

Git, GitHub, Docker, Raspberry Pi, VS Code Extension Development, PaddleOCR, Google Gemini API, OpenFDA Database.

Methodologies & Architecture

Agile Development, System Architecture, Modular Design, Data Processing Pipelines, Deterministic-first Architecture, Problem Solving, Cross-functional Collaboration.

Interests

Professional Interests

AI Agent Orchestration, Autonomous Systems, Web Development.

Projects

Pathfinder | Autonomous Tourism Agent

Summary

Developed a full-stack AI-powered autonomous tourism agent, leveraging a RAG system and advanced NLP for multilingual query resolution and efficient data processing.