Projects
Care AI Agent
An AI-powered virtual agent that handles customer care interactions end-to-end, from querying customer account data and service records to performing actions such as raising complaints, sending notifications, and escalating to live agents when needed.
Image AI
An AI-powered content moderation platform that validates and filters user-submitted images related to the Kumamon mascot character using multi-agent workflows.
Personal RAG
A single-user RAG application for chatting with your own notes, PDFs, bookmarks, and articles, with streamed answers and inline citations pointing to exact source chunks.
Delubyo
An interactive text adventure game set in the Philippines during a typhoon disaster, where your choices shape the narrative in an immersive storytelling experience inspired by Lifeline and Firewatch.
Local LLM Runtime
General-purpose infrastructure for running LLMs locally on a GPU, exposing a provider-swappable LLMClient abstraction and a FastAPI HTTP layer — designed as a reusable foundation for local RAG pipelines and agent projects.
Outlanders
A casual, adventure, and strategy-focused tower defense mobile video game that combines fun gameplay with learning about Philippine Literature, specifically Bugtong (Riddles), Alamat (Legends), Kwentong-bayan (Folktales), and Philippine Mythical Creatures
Work Experience
Full Stack LLM Developer
- Architected and designed AWS infrastructure for AI-assisted content moderation platform using ECS, ECR, S3, DynamoDB, VPC, ALB, CloudWatch, Secrets Manager, CloudFront, and configured IAM roles, permissions, and team development environments via AWS CodeCommit
- Developed FastAPI backend services and scalable database schemas for content upload and review workflows, and deployed both backend and frontend applications to ECS with comprehensive testing of deployments
- Developed a Python-Selenium web scraper to automate FAQ content extraction from E& B2B Portal across multiple support categories.
- Architected multi-step Agentic AI workflows on the amAIz and led KT sessions for fellow developers on how to utilize the amAIz for building and scaling agents.
- Defined and maintained data mappings across API schemas and technical documentation, supporting sprint planning.
- Coordinated and tracked System Integration Testing (SIT) across multiple backend systems, documenting blockers, escalating integration issues, and monitoring end-to-end API response coverage across testing layers.
- Diagnosed and resolved production and UAT defects in AI agent workflows, improving response accuracy, escalation handling, and system reliability
Python Backend Developer
- Architected data workflows with SQLAlchemy ORM and standardized API responses
- Implemented Alembic migrations for zero-downtime database updates
- Orchestrated Docker and Kubernetes containerization for multi-environment deployments
- Designed OAuth 2.0 authentication supporting Google and Sage Intacct integrations, with comprehensive token management, CSRF protection, and encrypted credential storage
- Implemented Casbin with combined ABAC and RBAC for multi-tenant hierarchies
Software Developer
- Built internal tools with Retool/Jira/Google Sheets/Discord Bot
- Automated Discord community engagement for 50K+ users via Python bots
- Provided API/GraphQL and game development support for developers and community
- Designed automated workflows and business processes via Jira/Retool integrations
- Provided technical support to community builders, ensuring timely delivery to engineering
Technical and Game Support
- Provided support to 30K+ players via Discord and Zendesk tickets
- Monitored and moderated Discord channels and servers, resolving technical issues and escalating complex cases to engineering teams
- Created and maintained a knowledge base for player issues and game guides
- Collaborated with cross-functional teams to improve game documentation and support workflows
About Me
Full Stack LLM Developer at Accenture, working at the intersection of backend engineering, cloud infrastructure, and applied AI. I architect and build AI-assisted systems end-to-end — from AWS infrastructure design and FastAPI backend services to agentic workflow development and frontend integration.
My recent work includes designing multi-agent AI pipelines on AWS Bedrock, building content moderation platforms with ECS, DynamoDB, and CloudFront, and engineering scalable backend services with Python, FastAPI, and PostgreSQL. I'm also comfortable on the frontend, building with React, TypeScript, and Next.js. AWS Certified AI Practitioner and Claude Certified Developer Foundation.
Before AI became central to my work, I spent two years at Axie Infinity building automation tools and supporting a 50K+ user community, then sharpened my backend skills at FiberFin with OAuth 2.0, multi-tenant RBAC, and zero-downtime database migrations. I care about clean, maintainable code and building things that actually solve problems.
Outside of work, I tinker with robotics and explore cybersecurity as side interests — breaking and building things for fun. I also spend a lot of time in Valorant and League of Legends. If you're curious about my gaming profiles, try typing komiwalnut on your keyboard. 🎮