The Business Challenge
The business situation the organisation faced before the project began.
Access to dermatologists is limited for many people in India, and most digital health tools are text-based and English-only. That leaves out users who are more comfortable speaking Hinglish, and single-model tools can give unreliable first assessments of skin conditions.
Why AI Was the Right Solution
Why AI was the right approach — and what alternatives were considered.
Specialist dermatology knowledge is scarce, and a symptom checklist cannot interpret images or natural speech. AI models can assess skin images and understand spoken descriptions, which makes an accessible first step possible where specialist time is limited.
How the Solution Was Delivered
Discovery, design, development, testing and rollout — the journey, not the tooling.
- 1Chose a problem with real need: limited access to dermatologists and English-only health tools.
- 2Scoped a version that could be built and deployed within 36 hours.
- 3Built a multi-model diagnosis flow instead of relying on a single model.
- 4Added a Hinglish voice interface so users can describe symptoms by speaking.
- 5Deployed the assistant and presented it, advancing to Round 2.
Key Technical & Architecture Decisions
Architecture, model selection, workflow and trade-offs.
Multiple models combined to give a more reliable assessment than one model alone.
Voice-first, Hinglish interaction to match how target users naturally communicate.
Positioning as a first step, not a replacement for a doctor.
A tightly scoped build so the product could be deployed within the 36-hour limit.
Challenges & How They Were Solved
Obstacles hit along the way and how they were overcome.
A very short timeline: handled by scoping tightly and focusing on the core flow.
Reliability of assessments: improved by combining several models.
Language accessibility: addressed with a Hinglish voice interface instead of English-only text.
User Adoption & Change Management
How users responded — training, change management and feedback loops.
The voice interface was the key to accessibility, since users can speak naturally instead of typing in English. Clear framing as guidance rather than diagnosis helps set the right expectations for a health tool.
Business Outcomes & Impact
The measurable outcomes from the underlying AI Deployment — with the story behind them.
Built and deployed in 36 hours; advanced to Round 2 of HackIndia Spark 4 2026.
The story behind the numbers
The assistant showed that AI skin-health guidance can be delivered in the language and mode many users prefer: spoken Hinglish rather than English text. Delivering a deployed, working product in 36 hours and reaching Round 2 demonstrated both the idea and the speed of execution.
Lessons Learned
What surprised the team, what worked well, and what would be done differently.
- Voice in the user's own language widens access to AI tools.
- Combining models improves reliability.
- Tight scoping makes ambitious hackathon builds deliverable.
- Health AI needs clear framing as guidance, not diagnosis.
Future Opportunities
Where this solution could go next.
Validation with dermatologists and clinical datasets.
Support for more Indian languages.
Referral to nearby clinics when a condition needs professional care.
Secure handling of health data for real-world use.
Professional Reflections
Guidance for another builder tackling a similar problem.
In health AI, accessibility and safety matter as much as model accuracy. Build for the language and interaction style users actually use, combine models where reliability matters, and be explicit that the tool supports rather than replaces professionals.