AI Systems Engineer · Agentic Platform Builder · Technical Writer
Building production AI systems, agentic workflows, and thoughtful software.
I write about AI architecture, platform engineering, product experiments, and the lessons that come from shipping real systems.
First
enterprise AI app deployed
Took an internal AI product from idea to production in a constrained environment.
Platform
architecture and systems work
Built reusable foundations for agentic workflows, integration paths, and operational visibility.
DocPipeline
product experimentation
Turned document extraction pain into a focused product and a clearer point of view on applied AI.
Explore
Architecture notes, systems work, and a few useful experiments.
Featured Writing
Writing on AI systems, architecture, and production tradeoffs.
Start here for the most useful notes.
How We Built Our First Enterprise AI Application
A field report on deploying an internal AI platform inside a real enterprise environment, from security reviews to operating trust.
Designing Reliable AI Pipelines: What Production Taught Us
Reliability in AI systems is mostly about interfaces, observability, fallback behavior, and good operational taste.
Why Most AI Systems Fail in Production
Five recurring failure modes that show up when promising AI demos meet compliance, cost, latency, and messy data.
Featured Project
DocPipeline shows how I approach production AI work.
Problem, architecture, impact.
DocPipeline
A document intelligence system for insurance workflows with OCR, LLM extraction, confidence checks, and human review.
- Problem
- Manual extraction was slow and inconsistent.
- Architecture
- A staged pipeline handled OCR, extraction, validation, and review.
- Impact
- Processing dropped from analyst-heavy work to a few minutes per document.
Goodies
Small tools, diagrams, and experimental scraps.
Compact, useful, and still in motion.
Latency Budget Scratchpad
A small planning aid for thinking through model, retrieval, orchestration, and human-review latency tradeoffs.
Open the lab
Agent Loop Diagrams
A growing set of annotated system diagrams for routing, retries, escalation, and human-in-the-loop patterns.
Open the lab
Prompt Evals Lab
A place for tiny experiments around prompt evaluation, scoring heuristics, and failure-case capture.
Open the lab