Justin · AI Platform & Agentic AI Engineer

Turn frontier AI intosystems that actually run.

Agent, RAG, and AI Platform engineering: architecture, evaluation, deployment, operations, and handoff.

View case studies Browse all projects ↗

Start with the question closest to yours

  1. 01 How do enterprise Agents stay controllable?Design evaluation, observability, permissions, and failure recovery into Agent workflows so they can be inspected and operated.
  2. 02 How can retrieval be made reliable?Turn scattered documents, permissions, and difficult queries into retrieval systems with measurable quality and traceable answers.
  3. 03 Need a technical talk or architecture exchange?Use an architecture review, PoC evaluation, or engineering talk to clarify the next decision for your team.

Flagship system: a measurable Agentic RAG pipeline

From 'a chatbot that can retrieve' to a measurable, deployable, and controllable enterprise knowledge base agent

Agentic RAG System

Problem
Mixed enterprise documents and unstable user phrasing made conventional RAG miss or mis-rank critical instructions.
Method
Built a controlled LangGraph workflow with rule-first routing, hybrid retrieval, context validation, and retries.
Outcome 98.0% weighted accuracy with 2.6-second average latency.
Read case study →

View all implementations

Latest research and engineering updates

Currently building

Enterprise Agent Technology Radar

Track models, agent harnesses, MCP, memory, evaluation, observability, security, and enterprise adoption, translating frontier signals into actionable engineering decisions.

Open the engineering log →
Browse all writing ↗

More production cases

PaddleOCR · YOLOv7 · Hospital Receipt Structuring · End-to-end Normalization

Receipt OCR API

Outcome Normalized formats from 5+ hospitals into API-ready JSON.
Read case study →

Speaking · Exchange · Collaboration

Bring frontier AI and engineering reality into the same conversation.

For teams and communities working through Agent, RAG, and AI Platform decisions.

Invite me to speak or connect