TechGPT

I build tech that makes healthcare work better for you.

I'm Rajesh, a healthcare technologist. I design AI agents, healthcare data platforms and EHR integrations that take work off clinicians' and patients' plates, and I help non-profits do the same on a small budget.

Get in touch Read my notes
  1. Patient apps
  2. EMR
  3. Claims
  4. Pharmacy
  5. AI agents

Three decades in technology. Over a decade in healthcare.

I've led engineering for organizations across the healthcare system, from the clinic to the pharmacy to the lab. The common thread: shipping software that real patients and clinicians use, and measuring whether it helped.

Providers
A patient portal for a leading academic medical centre.
Payers
Patient identity and point-of-care tools for a national insurer's primary-care clinics, covering millions of patients.
Pharmacy
The home-delivery experience for a national pharmacy benefit manager.
Life sciences
A partnering platform for the biotech industry, and enterprise AI platforms for global pharmaceutical companies.
Teams
Engineering teams of dozens, spread across North America and India, with the delivery, budget and client relationships that come with them.

What I build

  • AI agents for healthcare workflows

    Multi-agent systems that handle the back-and-forth of tasks like appointment scheduling, with the guardrails a regulated setting needs.

    Claude and Azure AI models, agent orchestration, Python

  • EHR and EMR integration

    Services that read and write clinical data reliably, including apps that run inside the EMR itself so clinicians don't have to switch screens.

    FHIR, HL7, EMR vendor APIs, Java and Spring Boot

  • Healthcare data platforms

    Pipelines, knowledge graphs and search that turn scattered clinical and claims data into something teams can act on.

    Databricks, Snowflake, Neo4j, pgvector

  • Quality review with AI

    Tools that review call-centre conversations against quality standards, so supervisors coach on patterns instead of sampling by hand.

    Speech-to-text, LLM evaluation, Python

How I work

The agents propose. The gates decide.

I use AI to build software, and I treat it as a fast, tireless colleague whose work still has to pass review. Architecture lives as code. Every pull request runs through automated evaluations before a person approves it. Every mistake becomes a rule, so the same one doesn't happen twice.

Working this way, small teams deliver what used to take large ones, without giving up the safety checks healthcare depends on.

I'm hands-on by conviction. When a release is at risk I write the code myself, and I'm direct about what AI can and can't do in a regulated environment.

TechGPT

A second opinion on your architecture and code.

TechGPT is an assistant I built on my own architecture practice. Ask it to explain a pattern, sketch a scalable design or look over a piece of code for defects. It has helped in more than 300 conversations so far.

Try TechGPT

Notes

Giving back

Good technology shouldn't be only for organizations with large budgets. In my own time, I'm glad to help with:

  • Technology and architecture advice for non-profits and community health organizations
  • AI-readiness conversations: where it can help, and what to watch for
  • Mentoring engineers moving into healthcare technology or AI
  • Talks for meetups and community events
  • Open, community-minded projects that could use another builder

Tell me a little about your organization and what you're trying to do.