AI / ML Platform Engineer
Xavi Ramirez
LLM systems, retrieval, evaluation and inference — in production, in regulated domains.
01 · About
I build and operate LLM and ML systems in production. Thirteen years of backend and platform engineering — Java, Spring Boot, Python, React, AWS — with the last four focused on applied AI in regulated insurance workflows.
02 · Writing
All posts →Notes on how AI systems actually work, and what it takes to run them in production.
Your own lesson pipeline: Efficacy & incrementality
Engagement counts what happened; efficacy asks whether it mattered. Section-level randomisation, a power calculation with the classroom clustering priced in, and the single modelling assumption that flips the same children and the same scores from “no detectable effect” to a number somebody would have published as a win.
Your own lesson pipeline: Publishing & fidelity
Diff, don’t republish. Re-importing a course strands sixty-three pieces of student work on the first night and a diff strands none — then read the funnel from what a teacher approved to what a child can actually use, and learn which of the losses are the gate working correctly.
Your own lesson pipeline: Guardrails, tenancy & deletion
A guardrail that runs only at generation is a default, not a guardrail. Publishability as a join the editor cannot route around, every teacher edit stored as an observation rather than an overwrite, and a deletion fan-out with one destination no amount of engineering can reach.
Your own lesson pipeline: Appropriateness as a computed property
“Age-appropriate” is four measurements wearing one word. Drive a rubric over real blocks and watch the same one pass in one district and hold in another — then see a same-family judge report 94% while the teachers who actually use the content rewrite six lessons in ten.
03 · Experience
Dreamcode.io — AI Platform Engineer
Guidewire Software — Senior Software Engineer, Submission Intake
Walmart — Senior Software Engineer
ScienceLogic — Senior Software Engineer
LatAm Autos, Devsu, Espaciolink — Software Engineer → Software Architect
B.Sc. Computer Science, Universidad de las Fuerzas Armadas (ESPE), Ecuador.
04 · Selected work
Retrieval architectures for insurance documents
in progressA benchmark of chunking, retrieval and reranking strategies over public P&C corpora — NAIC bulletins, specimen policy forms, ACORD specifications — measuring groundedness, latency and cost per query.
Chunking · Reranking · Groundedness · Cost per query
Role-based access control at retrieval time
in progressA demonstration of why filtering restricted content after generation leaks, and why the filter belongs in the retrieval query. Includes an MCP server and an exportable audit trail.
RBAC · MCP server · Audit trail
05 · Contact
Open to conversations about AI platform work. Reach me at me@xavier-ramirez.com.