Agentic AI Developer in Metro Manila
Den Jansen Flores is an agentic AI engineer and full-stack developer in Metro Manila. He builds AI features that perform bounded work inside real applications: calling typed tools, retrieving grounded context, returning schema-validated output and carrying state across guarded steps. He is available for freelance, contract and full-time work.
What does agentic AI engineering include?
The model is one component. The engineering value sits around it: authority, tools, state, evaluation, observability, recovery and cost control.
Typed tool workflows
Models call explicit functions with validated arguments instead of improvising actions against production systems.
Grounded retrieval
Relevant business context is retrieved, scoped and attached to the task so answers can be traced to known data.
Guarded state and recovery
Multi-step work keeps an inspectable state, names failure paths and requires approval where an automated action would carry real consequence.
Data and analysis workflows
Ingestion, normalization, validation and analysis pipelines that turn uploaded or operational data into dependable decision support.
When should a team use an AI agent?
- A repeatable task requires several tools or decisions, not one generated paragraph.
- The system can define what the model may read, call, change and escalate.
- A human needs evidence, approval or recovery controls around automated work.
- The business can measure success, latency and cost per completed task.
Projects connected to this service
The published work covers AI-assisted business analysis, assessment workflows and operational software prepared for controlled automation. Claims stay tied to case studies, source status and dated deployment checks.
Data Analysis Platform
AI data platform case study: normalise business uploads, constrain OpenAI output and render validated analysis in a dashboard.
Career Path
Student assessment case study with explained career recommendations, role-based access and MongoDB support for changing cohort formats.
ClinicFlow
Clinic operations case study covering doctor availability, appointments, patient records and role-separated staff and patient portals.
How the engagement moves
Define the bounded job
State the decision or task, available evidence, allowed tools and the exact conditions that require a human.
Design the control layer
Specify schemas, permissions, retries, timeouts, budgets and audit records before connecting a model.
Evaluate real failure modes
Test groundedness, tool selection, malformed output, unavailable dependencies and recovery rather than judging one polished demo.
Ship with visibility
Expose status, evidence and owner controls so the system can be trusted, corrected and stopped.
Questions before contact
Is Den available as a freelance agentic AI developer in Metro Manila?
Yes. Den is based in Metro Manila and is available for freelance, contract and full-time agentic AI or full-stack engagements, including remote work with teams outside the Philippines.
What is the difference between an AI agent and a chatbot?
A chatbot mainly returns language. An agent performs a bounded workflow: it may retrieve context, call typed tools, maintain state and request approval. Those actions require permissions, validation, observability and recovery controls.
Which AI platforms can he work with?
His application work uses OpenAI and Claude APIs, tool calling, retrieval and schema-constrained output. Provider choice follows accuracy, latency, privacy, cost and deployment requirements.
Can he add AI to an existing SaaS product?
Yes, when the product has a clear task and data boundary. He can map the workflow, add the model behind typed interfaces and preserve human approval for actions that should not run unattended.
Bring the workflow that needs an owner.
Send the current process, the failure it creates and the deadline that matters. You will get a direct response from Den, usually within one business day on UTC+8.
Email Den