By Tawhidur Rahman Laskar
Founder & Project Leader, Hippocrates India
What happens when a student interested in software engineering decides to explore one of the most challenging applications of artificial intelligence: healthcare?
For me, that question has become a real building project.
I'm working on HIPPOCRATES, an AI Clinical Platform under Hippocrates India. The goal is to explore how AI can make healthcare information easier to understand and support clinical workflows without replacing qualified healthcare professionals.
Project website: https://www.hippocrates.in/
The problem I'm exploring
AI applications are becoming easier to build, but healthcare presents a different set of challenges.
A useful healthcare tool needs to do more than generate fluent answers. It must account for uncertainty, communicate limitations, handle sensitive information responsibly, and avoid presenting unverified output as established medical fact.
I wanted to explore how these challenges could shape the design of an AI-powered healthcare platform.
Two products, two different users
HIPPOCRATES is organised around two products.
- Aegis — Patient-focused AI assistance
Aegis is being developed to help people better understand their health information.
The intended capabilities include:
Explaining medical report terminology in plain language.
Helping users organise symptoms and describe their concerns.
Preparing questions to discuss with healthcare professionals.
Making complex health information easier to navigate.
The central design challenge is accessibility without creating false confidence. An AI explanation should not be confused with a confirmed diagnosis or personalised medical advice from a qualified professional.
- Oracle — AI copilot for clinicians
Oracle focuses on supporting healthcare professionals with information-intensive tasks.
Its intended areas of support include:
Clinical research and information review.
Structured clinical documentation, including SOAP notes.
Differential-diagnosis support and clinical checklists.
Guideline cross-referencing and evidence-informed decision support.
The objective is to support professional workflows, not automate away clinical responsibility.
What makes healthcare AI challenging?
Building an interface around a language model is only one part of the work. The harder questions involve how the system should behave when the information is incomplete, ambiguous, or potentially high-risk.
Some of the questions I'm exploring are:
How should uncertainty be communicated?
Users need to understand the difference between an explanation, a possibility, and a verified clinical finding.
How should safety influence product design?
High-risk situations require appropriate escalation rather than an ordinary conversational response. AI outputs also need to be treated as fallible and checked appropriately.
How can complex information become accessible?
Medical terminology can create a barrier between information and understanding. Good product design should reduce that barrier without oversimplifying important details.
How should humans remain in control?
Healthcare professionals must retain responsibility for clinical decisions. AI should assist with information and workflows rather than independently determine patient care.
These are not problems that can be solved simply by writing a better prompt. They require thoughtful product design, evaluation, and ongoing improvement.
Building while learning
I'm still learning and developing my skills in software engineering, artificial intelligence, and product development.
Working on HIPPOCRATES gives me an opportunity to think beyond getting a feature to work. It pushes me to consider what happens when real people interact with a system, how errors might affect them, and what responsible design should look like.
I believe this is an important part of becoming a software engineer: learning to build systems that are not only functional, but also understandable, useful, and designed with their limitations in mind.
The project is ongoing, and its capabilities need continued testing and validation. I see this as a long-term engineering journey rather than a finished solution to healthcare's challenges.
What's next?
My focus is on continuing to develop Aegis and Oracle, improving the user experience, and exploring responsible ways to make healthcare information more accessible.
I'm particularly interested in connecting with developers, healthcare professionals, researchers, and other builders working on responsible AI applications.
If you're exploring similar problems, I'd love to hear what you're building and what challenges you've encountered.
Explore the project: https://www.hippocrates.in/
Connect with me: https://www.linkedin.com/in/tawhidlaskar/
HIPPOCRATES is an evolving healthcare technology project. Its tools are intended to support health-information understanding and professional workflows, not replace qualified medical care or professional clinical judgement.
About me: I'm Tawhidur Rahman Laskar, a student innovator, aspiring software engineer, and Founder & Project Leader of Hippocrates India.
Top comments (0)