nyra Brain Days 2026
and present
crews of 4 or 5 people
per team
01The Plan
Day 1 is about finding the right idea and getting started. Day 2 is heads-down building. Day 3 is showtime.
Day 1 — Wednesday, April 15 IDEATE + START
Day 2 — Thursday, April 16 BUILD
Day 3 — Friday, April 17 SHOWTIME
02The Basics
This is a 3-day hackathon. You'll have 2 full days to build and 1 day to demonstrate what you've created. The goal is not to deliver production code — it's to explore what's possible.
Core Constraints
- Every person MUST use an AI tool for one of their tasks. This isn't optional. Bonus points if AI is doing something genuinely hard or surprising.
- Creativity first, roadmap alignment optional. You're not constrained by current product priorities. Go bold.
- Demo-ready > production-ready. Make something compelling enough to demo, not to ship next week.
- Small scope wins. A focused idea executed well beats an ambitious idea half-built.
Brain Days
- A space to explore ideas outside the roadmap
- Cross-functional — every role contributes
- A pitch package, not just a code demo
- Small, focused scope preferred
A sprint
- Not a sprint to ship features
- Not just for engineers — your domain knowledge is your competitive advantage
- Not a solo project — you're paired for a reason
03Team
Unlock your team
04Choose Your Path
Pick one track on Day 1 during the brainstorming session (10:00–13:00). The examples below are just starting points — you're encouraged to come up with your own ideas within each track.
Product Improvement
Litmus test: Could this realistically ship within 3 months?
Make nyra better for therapists or patients using AI. These are real problems our users face every day.
Suggestions to get you started:
- myReha “Reaching your personalized goals”: A conversational interface during onboarding that asks what the overall goal is that the patient tries to achieve. An AI transforms this into quantifiable goals, which are then tracked, re-evaluated and measured over time.
- nyra insights “Ask anything”: Connect BigQuery functionality in insights with a chatbot, which can produce charts and show data about patients.
- Family dashboard: Caregivers (spouses, children) get a weekly summary: “Mom completed 4 of 5 sessions this week. Her naming accuracy improved 12%. Here’s what you can practice together at dinner.” Turn the family into a therapy multiplier.
Internal Tooling
Litmus test: Would this actually save someone time, today?
Kill a painful internal workflow with AI. Think about something that took you 2+ hours last week that could be automated.
Suggestions to get you started:
- DRV reporting: Create beautiful reports for different clinics that can be shared via passwords.
- QMS check: Search and cross-reference our regulatory guidelines with the actual implementation.
- Exercise preview: Build a Claude project that brainstorms exercise descriptions and then visualizes a prototype for fast iteration.
Moonshot
Litmus test: Is this exciting enough that you'd pitch it to investors?
Forget the roadmap. Build something that would be amazing if it existed, constrained only by: it uses AI, it benefits nyra, and you can demo it in 2.5 days.
Suggestions to get you started:
- nyra insights / myReha scribe solution: Automated session note generation from recordings. SLP speaks to patient, AI generates the clinical note.
- myReha “Copilot”: An AI avatar that helps out on every exercise in myReha — it analyzes the scene visually, gives constructive feedback on how to solve it, and motivates along the way.
- Patient self-therapy companion: An AI voice agent that guides patients through home exercises between sessions. Not a chatbot, but an actual voice conversation: “Okay Mario, let’s try the naming exercise. I’ll show you a picture, you tell me what it is. Ready?” Tracks performance, reports back to therapist.
05The 20-Minute Pitch
Every team member should present their part.
| Time | Section | Who presents | What to show |
|---|---|---|---|
| 4 min | Problem & Clinical Context | Domain expert | What's broken? Why does it matter? Real pain, real user, real context. |
| 6 min | Live Demo | Domain expert + Developer | Show the prototype working. |
| 4 min | GTM & Business Case | Sales / BD / Growth | Depending on chosen track - market opportunity, one-pager, internal expedited processes, ... |
| 4 min | How We Built It + AI | Developer / ML Engineer | Technical approach, AI tools used, key decisions |
| 2 min | Q&A | Whole team | Designate one person to field questions. Be concise. |
06Common Pitfalls
Every hackathon has the same failure modes. Here's how to avoid them.
Overscoping
The #1 project killer. Build the smallest thing that demonstrates the idea. You can always add features — you can't add time.
Architecture before user story
Start with the user story. What are they trying to do? What does success look like? Then build the minimum tech to get there.
Not pivoting early
Your initial idea will hit a wall. That's normal. If it's not working by Day 2 morning, change course. The "Best Pivot" prize exists for a reason.
Not asking for help
Founders are floating for a reason. Stuck on AI integration? Scope unclear? Grab a founder. You have world-class help available — use it.
Build the demo first
Teams that start with "how do we show it?" and work backward produce better results. You'll cut scope naturally. Demo-first, infrastructure-never.
07Judging & Prizes
Judged by group poll. Focus on impact and creativity — you don't need production code.
| Weight | Criterion | What the judges look for |
|---|---|---|
| 30% | Business Impact / User Value | Does this solve a real problem for real people? |
| 25% | Innovation & Creativity | Would this have been obvious before the hackathon? |
| 20% | Demo Quality & Presentation | Compelling story, clear demo, engaged audience? |
| 15% | Technical Execution | Does the prototype work? Could it be extended? |
| 10% | Use of AI | Is AI deeply integrated, or tacked on? |
Prizes
Best Overall
Highest cumulative score
Most Likely to Ship
Could ship within 3 months
08Your Arsenal
What to install and what to think about before you arrive.
Build with
| v0 | AI design & React component generation |
| Lovable | Rapid web app prototyping with AI |
| Claude | API access via Claude Code CLI |
| Company Info Hub | Full product, clinical, and ops docs |
| Pre-configured MCPs | Ready-to-use integrations |
Checklist
- Install Claude Desktop app. Make sure you can open a repo and run code.
- Review your teammates. What does each person do? Who are the domain experts?
- Brainstorm 2–3 project ideas. Don't commit — just think about what excites you.
- Read the context about nyra. Know what we do, who our users are, what frustrates them.
- Come with an open mind. Initial ideas often pivot. That's expected.