Security
Least privilege and controlled access from the first design conversation.
AI / ML engineer
I founded Caelara to bring enterprise AI delivery to education. I help universities, medical schools, and families design systems that can be audited, operated, and trusted.
10+
Years in engineering
6
Years in data science, ML & GenAI
3
Clouds we build on
Education
Primary focus
I have spent more than ten years as an engineer. For the last six of those years I have delivered data science, machine learning, and generative AI as products and services inside enterprises.
That work taught a simple discipline. Models are easy to demo and hard to operate. Security, privacy, compliance, reliability, and observability are the product. If a system cannot be audited, isolated, monitored, and owned, it is not finished.
I build on AWS, Azure, and Google Cloud. Institutions rarely live in one provider. Identity, networking, secrets, logging, and cost are part of the engagement — not a later IT ticket.
I now focus on education because I care about learning, and because the sector is under-served. AI investment follows profit. Classrooms, medical schools, and homeschool families inherit tools that were never designed for them. Caelara exists to do that work.
Least privilege and controlled access from the first design conversation.
Student and family data stay minimized, isolated, and explicitly retained.
Built with FERPA, accessibility, and institutional review in mind.
Production systems with clear failure modes, fallbacks, and ownership.
Logs, traces, and evaluations so you can see what ran, why it failed, and what it cost.
Production delivery on AWS, Azure, Google Cloud — including the identity, networking, and operations work that makes a model usable inside an institution.
AI delivery for a medical school, including accessibility tooling and faculty discovery work.
AI consulting for a research university that needs production systems, not slideware.
Direct work with families who teach at home, starting from real constraints.
Tell us about the program, the constraint, and what done looks like.