• Contact Info
  • +63 970 582 6674
  • james.ryan.pro81@gmail.com

Home

Photo

I'm James Ryan

About me

/

About Me

Hello, I'm James Ryan - AI Engineer

Degree: Bachelor of Science (BS)

Email: james.ryan.pro81@gmail.com

Phone: +63 970 582 6674

6

Years Experience

4

Companies

5

Certifications

3

Industries

Resume

My Resume

AI Engineer with around six years of experience building and shipping production software, the last few focused on LLM applications. I design retrieval augmented generation pipelines, build AI agents that handle multi-step tasks, and integrate generative AI into existing products without breaking what already works. Most of my recent work has been in healthcare, fintech, and billing - where accuracy and traceability matter - so I care a lot about evaluation, guardrails, and keeping cost and latency under control.

Skill

LLM App Development / RAG / AI Agents
100%
Agent Orchestration (LangGraph / LangChain)
95%
Prompt Engineering / Structured Output / Function Calling
95%
OpenAI / Anthropic / AWS Bedrock APIs
95%
Vector Databases (pgvector / Pinecone)
90%
LLM Evaluation / Ragas / LangSmith
90%
Workflow Automation (n8n) / AI Pipelines
90%
Python / TypeScript / SQL
95%
FastAPI / Node.js / REST APIs
90%
PostgreSQL / MySQL / MongoDB
90%
Redis / Elasticsearch
85%
AWS (Bedrock / SageMaker / Lambda)
85%
Docker / CI/CD (GitHub Actions)
85%
Datadog / Sentry / Observability
80%

Experience

  • Senior AI Engineer
    Collectly · United States, Remote06/2025 – Present
    • Lead AI engineering on a patient billing platform, building agentic workflows that read statements, answer balance and insurance questions, and route edge cases to staff - cutting manual review on routine tickets by ~45%.
    • Designed a RAG layer over payer policies, billing codes, and account history so the assistant gives grounded answers with citations, raising answer accuracy in internal evals from the high 70s to ~92%.
    • Implemented a labeled regression set and automated scoring in CI so prompt and model changes are checked before release.
    • Tuned retrieval, caching, and model routing to send simpler calls to smaller models, dropping average response latency by ~35% and trimming monthly LLM spend by ~30%.
    • Mentor two engineers on prompt design, eval practices, and safe rollout patterns for AI features.
  • AI Engineer
    Helicap · Singapore, Remote01/2024 – 05/2025
    • Built LLM tools for a lending and credit investment team, including a document assistant that extracts figures and terms from financial PDFs - reducing manual data entry time per deal by ~60%.
    • Shipped a RAG-based research assistant over company filings and internal memos using pgvector and FastAPI, letting analysts surface context in seconds instead of digging through folders.
    • Added agent-style tool calling so the assistant could query internal APIs and databases for live numbers, with guardrails to prevent returning unverified figures.
    • Set up tracing and cost dashboards for every LLM call to surface token usage, latency, and failure rates per feature.
  • Software Engineer
    Alliance Software Inc. · Manila, Philippines05/2022 – 12/2023
    • Built and maintained backend services in Python and Node.js for client-facing web products, owning API design, database work, and third-party integrations.
    • Delivered an early generative AI feature that drafted customer support replies from a knowledge base, helping a pilot team cut average first response time by ~40%.
    • Improved a slow reporting endpoint by reworking queries and adding caching, reducing a common request from several seconds to under 400ms.
    • Added a test suite and GitHub Actions pipeline to a service that had none, noticeably reducing bugs reaching production.
  • Junior Software Engineer
    VintaSoft Inc. · Zamboanga City, Philippines08/2021 – 04/2022
    • Wrote backend features and bug fixes in Python for internal business tools, learning the codebase through tickets and code review.
    • Built REST endpoints and database queries for a document processing tool and added input validation that reduced bad data getting saved.
    • Wrote unit tests and helped move part of the team toward consistent testing practices.

Education

  • WMSU
    Western Mindanao State University
    2017 – 2021 · Zamboanga City, Philippines

    Bachelor of Science in Computer Science

    Graduation

Certifications

Deep Learning
Deep Learning
TestDome
Machine Learning
Machine Learning
TestDome
AWS Certified Developer
AWS Certified Developer
TestDome
Problem Solving
Problem Solving
HackerRank
Python
Python
HackerRank

My Services

My Services

Services I Offer To My Clients

LLM Application Development

I build production-grade LLM applications using OpenAI, Anthropic, and AWS Bedrock APIs - with a focus on accuracy, cost control, and safe integration into existing products.

RAG Pipeline Design

I design retrieval augmented generation pipelines with vector databases (pgvector, Pinecone), semantic and hybrid search, chunking strategies, and reranking to maximize answer grounding and traceability.

AI Agent Development

I build multi-step AI agents with tool calling and agent orchestration (LangGraph, LangChain) that automate complex workflows and integrate with internal APIs and databases.

LLM Evaluation & Guardrails

I set up evaluation frameworks with Ragas and LangSmith, build labeled regression sets, and add hallucination and safety checks so model changes are caught before they reach production.

AI Workflow Automation

I automate data ingestion, indexing, and AI-driven processing pipelines using n8n, webhooks, and API integrations to reduce manual work and keep knowledge bases fresh.

Cloud & Deployment

I deploy AI systems on AWS (Bedrock, SageMaker, Lambda) using Docker and GitHub Actions CI/CD, with observability via Datadog and Sentry to monitor cost, latency, and failure rates per feature.

Portfolio

My Projects

Computer vision and deep learning projects

Object Detection
Object Detection
Computer Vision
AI Conversational Avatar
HeyGen · WebRTC · LLM
AI Image-to-Video Generation Pipeline
ComfyUI · Wan Model
AI Virtual Staging for Real Estate
AI Image Generation · Real Estate
AI Customer Support Agent
Gmail · Gemini · LLM
Workflow Automation & AI Agent Systems
Workflow Automation & AI Agent Systems
n8n · LLM · No-Code
LinkedIn Lead Generation Automation
Make.com · Airtable · PhantomBuster
OCR & Scene Text Recognition System
OCR & Scene Text Recognition System
Computer Vision · OCR
LeaseFlow AI
Voice AI · LLM Agents · Real Estate

Get in Touch

Get In Touch

Feel Free To Contact Me Any Time!

Let’s Get Work With Me

I always provide technical, creative and business insights to all projects I work on. With my experience and enthusiasm, I am sure I would be a great addition to your team and be a key player in making your project successful!