Jarvis V1, my home medical assistant

• 

  Jarvis

Inspiration – Iron Man The idea for Jarvis was inspired by Iron Man’s AI assistant. I believe every individual—and every home—should have access to a personal AI system to help a family. Think of the human body: we inhabit it, but we don't consciously manage every function. That’s the job of the parasympathetic nervous system, which quietly ensures balance so we can focus on higher-level tasks. Why should the homes we live in be any different? A home AI should quietly manage our environment in the background, allowing us to live with greater ease, health, and intelligence.

What It Does Jarvis is the first version of my personal home AI. Its initial role is to manage medical inventory within our home, starting with our medicines. My wife’s father is a physician, and over the years we’ve accumulated around 2,000 individual medications spanning 150 distinct types. As an Indian household, we follow both Western and Eastern (Ayurvedic) medical practices. Jarvis has been trained on a curated knowledge base that includes both systems, along with a detailed inventory of our home’s medicine collection.

How Jarvis works: When I encounter everyday health concerns—like headaches, allergies, body pain, fever, or minor infections—Jarvis identifies the appropriate medical molecules typically prescribed. It then cross-references these with our home inventory to provide a list of relevant medicines, including dosage and precise location. It also searches WebMD and trusted Ayurvedic resources to recommend holistic alternatives. Finally, Jarvis helps evaluate the best course of action and emails a plan, all while grounding every decision in reliable, vetted sources using Retrieval-Augmented Generation (RAG).

How We Built It I’m not a programmer by background. I began learning to code just 3–4 months ago during a sabbatical, driven by a deep interest in bringing Jarvis to life. My journey started with Python, but I soon discovered VectorShift, an agentic development platform, which accelerated my progress. I used VectorShift to build the Jarvis agent and leveraged multimodal AI to construct our medicine database. ChatGPT was a fantastic tutor through my process - to understand core AI and agent concepts, and to troubleshoot bottlenecks along the way.

Challenges We Ran Into One of the biggest challenges was the absence of a clear, curated pathway to learn AI. I had to navigate a fragmented learning landscape—piecing together tutorials, documentation, and forums. It was a process of trial and error, full of stumbles and learnings, until I eventually found the right combination of resources and tools.

Accomplishments We’re Proud Of Jarvis is fully functional and live in our home. It actively supports our household’s medical needs and demonstrates that meaningful AI applications can be built even without a formal technical background.

What We Learned This is just the beginning. Jarvis has opened my eyes to the potential of deeply personalized, useful home AI systems. There is so much more it can learn to do—and I now have a foundation to keep evolving it.

What’s Next for Jarvis The next milestone for Jarvis V1 is to become our home pantry and food assistant. The goal is to apply the same intelligence and structure to managing our kitchen inventory—helping us plan meals, reduce waste, and maintain better nutrition at home.

Built With

• database
• google
• integrations
• langchain
• vector
• vectorshift