ABOUT ME
Hey, I'm Marvin.
AI Engineer with a background in SQL Server database systems and production infrastructure.
Before transitioning into AI, I spent several years working with Microsoft SQL Server in production environments, where I developed experience in database administration, automation, monitoring, performance tuning, backup and recovery, and operational reliability.
That background influences how I approach AI engineering today—building systems that emphasize reliability, maintainability, observability, and scalability rather than simple demos.
THE MAP
What would you like to know about me?
Choose a path into my background, projects, ideas, or the ways we can work together.
01 · PROFILE + SKILLS
Who am I?
An AI engineer focused on turning ambitious ideas into useful systems—fundamentals first, frameworks second. These are the capabilities I take from idea to production.
CORE
Python + TypeScript
The languages I use to build AI services, interfaces, and production workflows.AI SYSTEMS
LLM applications
Prompt design, structured outputs, tool calling, and reliable model integrations.RETRIEVAL
RAG + search
Grounded answers built with thoughtful chunking, retrieval, reranking, and evaluation.AGENTS
Agentic workflows
Human-aware systems that plan, use tools, recover from failure, and stay observable.PROTOCOLS
MCP + integrations
Connecting intelligent applications to the real tools and data they need.DELIVERY
Product engineering
Turning uncertain ideas into focused, testable software people can actually use.02 · SELECTED WORK
What have I built?
Real systems, clearly explained. Project links, screenshots, repositories, and verified results will replace these structured placeholders.
PROJECT · CASE STUDY COMING
Knowledge Retrieval Engine
A source-aware research system built around hybrid retrieval, grounded synthesis, and measurable relevance.RAG · PYTHON · EVALSPROJECT · CASE STUDY COMING
Operations Agent
A tool-using agent that turns messy requests into visible workflows with approvals at the right moments.AGENTS · MCP · TOOL CALLINGPROJECT · CASE STUDY COMING
Unified Model Gateway
A resilient model interface with routing, usage controls, structured telemetry, and provider fallbacks.TYPESCRIPT · APIS · OBSERVABILITYVisitors will be able to open the demo, read the case study, and inspect the code.
→03 · NOTES + BLOG
How do I think?
The blog will organize practical lessons from building AI systems into useful trails—not disconnected posts.
FIELD NOTE · PLANNED
Retrieval quality starts before the vector database
Why content boundaries and evaluation questions shape everything downstream.LAB NOTE · PLANNED
What makes an agent trustworthy?
The product decisions that matter more than choosing an orchestration framework.SYSTEM NOTE · PLANNED
The small model gateway I keep rebuilding
A practical architecture for routing, fallbacks, usage data, and provider independence.04 · CONTACT
Let's build something useful.
Have an AI problem, a role, or an idea worth testing? Send Marvin a note.
jbmarvin21@gmail.com →