I design agentic AI, data platforms, and the operating practices that make them useful in real organizations—from retrieval and orchestration to monitoring, governance, and adoption. My work lives where emerging AI meets enterprise reality.
Build it.
Make it work.
Teach others.
A recurring pattern across conversational AI, data engineering, MLOps, and agentic systems.
I have spent my career turning unfamiliar, high-friction problems into working systems. I began as the sole full-stack developer behind a banking chatbot, moved through cloud-scale data engineering, built the operating discipline for production ML and LLM systems, and now develop multi-agent experiences for enterprise analytics and reporting.
The common thread is not a single tool. It is ownership: learning the domain, designing the workflow, integrating the system, watching how people use it, and creating the standards and shared knowledge needed to keep it useful.
I am comfortable being the first person in the room to turn an emerging capability into a practical product or operating model.
I connect AI interfaces to data pipelines, cloud infrastructure, monitoring, governance, and the people who operate them.
I turn certifications and project lessons into bootcamps, review sessions, lectures, and communities of practice.
Selected examples are described at a level appropriate for public discussion. The emphasis is on the problem, system design, and operational impact.
Pioneered an agentic AI initiative in JPMorganChase Philippines for analytics and reporting. The system orchestrates a retrieval agent for report requests and internal documentation with an agent that can query and analyze data using SQL.
Impact: The initiative earned a TALA Award for Operational Excellence within my first year at the firm.
Exploring agentic GraphRAG with TigerGraph and developing interactive agent experiences with AG-UI and CopilotKit—an evolution from static dashboard consumption toward reports users can question, shape, and create.
Direction: Combine graph context, analytics, and generative interfaces without losing reviewability.
Created MLOps policies and procedures for model deployment, monitoring, access, infrastructure, data drift, and score drift. Built Airflow workflows and monitoring dashboards, supported DataRobot operations, and later created an LLM Operational Framework covering deployment, testing, monitoring, responsible use, and governance.
Impact: Helped operationalize a growing production estate that included active monitoring for at least 70 ML models.
Served as the sole full-stack developer for the early ESTA conversational banking experience, beginning with credit-card applications and expanding into servicing, auto loans, collections, and more than 20 automation initiatives.
Impact: Built the interface, backend integrations, early user acquisition experiments, intent routing, and image validation as adoption grew.
Studied AWS EMR cost and performance behavior to support cluster-sizing decisions for large transformation workloads, then implemented metadata-driven orchestration with Lambda and Spark. Later helped modernize an enterprise communications pipeline using Kafka, Azure Data Lake, Databricks, and a medallion architecture.
Impact: The Azure proof of concept moved pipeline processing from hours to minutes and progressed through development and UAT.
Built a convolutional neural network for Filipino tweet sentiment analysis using distant supervision, recognized as Best Undergraduate Special Problem. Later worked with the DOST and Philippine Genome Center on molecular modeling and dynamics simulation of integrin heterodimers.
Recognition: Both research efforts received invitations for conference poster presentation.
The career story is a progression from building the interface, to engineering the data underneath it, to operating AI at scale, to designing agentic systems and technical communities.
JPMorganChase Philippines
Agentic AI Engineer
Quant Analytics Associate
Build + lead: Multi-agent knowledge and analytics systems, GraphRAG exploration, interactive reporting experiences, a TALA Award for Operational Excellence, and the Ignite Data Science, ML, and AI Community of Practice.
GCash
Lead ML Operations Engineer
Operate + govern: MLOps policies, model monitoring, Airflow orchestration, DataRobot operations, dashboards, and an enterprise LLM Operational Framework.
Accenture
Data Engineering Senior Analyst
Scale + teach: AWS workload orchestration, Azure data-platform modernization, extensive cloud and data certifications, Apache Spark bootcamps, and certification review sessions.
EastWest Bank
Assistant Manager
Digital Systems Development
Build from zero: Full-stack conversational banking, backend automation, NLP and vision capabilities, digital acquisition experiments, and the expansion of ESTA into a broader service channel.
UP Mindanao + PGC
Computer Science Researcher
Bioinformatics Intern
Learn through inquiry: Filipino-language NLP, molecular simulation, academic recognition, and the research habits that still shape how I experiment with new systems.
The model matters, but so do the workflow, data, interfaces, controls, incentives, and people around it.
Map the actual decision, handoff, bottleneck, and user expectation before choosing the AI pattern.
Enterprise users need grounded outputs, visible assumptions, and a practical way to check the system.
Monitoring, ownership, escalation, access, and change management should exist before the launch celebration.
A strong system leaves behind stronger engineers, operators, and communities—not just code.
Teaching is not a side note in my career. It is how I consolidate new knowledge, raise the capability of a team, and connect specialists who would otherwise stay in separate rooms.
Launched a Data Science, Machine Learning, and AI community in JPMorganChase Philippines, bringing practitioners together for shared learning and beginning with an introduction to agentic AI.
The technology changes by problem. The durable capabilities are system design, data fluency, production judgment, and the ability to learn a new domain quickly.