Handbook - OpenAI Realtime Voice x Reasoning Hackathon (AI Tinkerers)
AI Tinkerers - New York City
Sign in

OpenAI Realtime Voice x Reasoning Hackathon (AI Tinkerers)

Handbook for the OpenAI Realtime Voice x Reasoning Hackathon (AI Tinkerers) hackathon.

Sign in
DevRel Lead - Restate
Restate is building durable infrastructure for AI agents and backends. We're looking to hire our first DevRel Lead to spread the word and bring Restate to developers.
Founding Applied AI Lead - Paxos Health
Stanford-founded Seed-stage healthcare AI startup with >$5M in VC funding and AI agents deployed in production with customers. You will build/own our agent patterns, prompt chains, structured outputs, evals, etc. for documentation, voice AI, and computer-use agents. You'll sit at the intersection of product, engineering, ops, and AI at a fast-moving startup. The best description of the job is in our official job posts (please apply there): - NYC: https://jobs.ashbyhq.com/Pear-VC/489d22ab-558f-4702-975b-e44c4d413205 - Toronto: https://jobs.ashbyhq.com/Pear-VC/2a4a4733-3471-4718-b11e-dea1e70a27bd
All Jobs
Final round winners have been announced. View Results

Handbook

##WELCOME!
Thank you for participating in the OpenAI Realtime Voice x Reasoning Hackathon, sponsored by OpenAI at Betaworks. This guide will help you get started with building cutting-edge realtime voice and reasoning experiences using the latest OpenAI technologies.
The goals of this event are:

Build cutting-edge realtime voice and reasoning experiences with AI that respond at the speed of thought
Create applications using the Realtime API and o1
Learn from and connect with industry leaders and OpenAI engineers
Transform realtime voice AI from demo to reality with new architectures
Push the boundaries of instantaneous AI interaction

###How Judging Will Work
The preliminary round of judging will evaluate projects based on four key criteria:

Running Code (25%)
Innovation & Creativity (25%)
Real-world Impact (25%)
Theme Alignment (25%)

The preliminary round of judging will take place in two rooms in divide and conquer style. Each team will play the video they submitted and judges will have a chance to ask a couple of questions and then they will place their votes. This will maximize the number of votes per team and give a chance for teams to present in front of a large number of attendees

Based on this, the judges will tally the results and select 7 finalists. The 7 finalists will present in front of the all attendees, judges and guests on the main stage in a live demo format – 5 minutes each.
###GETTING STARTED WITH REALTIME VOICE X REASONING

####Core Components
To build a realtime voice x reasoning system, you’ll typically need these components:

  1. Voice Processing Layer

-Real-time speech-to-text processing
-Voice activity detection
-Sub-second response capabilities

  1. Reasoning Layer

-Instant inference and decision making
-Context management
-Multi-step reasoning pipelines

  1. Integration Layer

-WebSocket or real-time communication protocols
-State management
-Error handling and recovery

####Key Principles

When designing realtime voice x reasoning systems, consider these principles:

Speed: Aim for sub-second response times
Reliability: Handle edge cases and errors gracefully
Context: Maintain conversation state effectively
Scalability: Design for production-ready performance
User Experience: Create natural, fluid interactions

###TOOLS & RESOURCES FROM SPONSORS

#####OpenAI

Access to Realtime API and o1
Documentation and implementation guides
Support from OpenAI engineers

#####Nebius

Direct access to NVIDIA GPUs (H100s, H200s)
InfiniBand networking
Managed Kubernetes and MLOps tools

#####Kamiwaza AI

Enterprise GenAI Engine
RAG tools and local inference setup
Location-aware inference mesh

#####Neon

Serverless Postgres platform
Autoscaling and instant branching
AI-ready capabilities

###PROJECT IDEAS

Here are some potential projects to consider:

  1. Real-time Voice Assistant

Sub-second response times
Context-aware conversations
Multi-turn reasoning capabilities

  1. Interactive Learning Platform

Voice-based educational experiences
Real-time feedback and corrections
Adaptive difficulty based on user performance

  1. Voice-Enabled Analytics Dashboard

Natural language queries with instant responses
Real-time data visualization updates
Complex reasoning for data analysis

  1. Collaborative Voice Workspace

Multi-user voice interactions
Real-time decision support
Shared context management

  1. Accessibility Tool

Real-time speech-to-text with reasoning
Context-aware assistance
Adaptive user interface

###RULES
####Must Have:

Include real-time voice interaction
Demonstrate reasoning capabilities
Have working code (not just mockups)
Participate in person - no remote teams allowed

####Don’t:

Submit pre-built projects
Create harmful applications
Share other teams’ private code or ideas
Use copyrighted material without permission

###EVENT LOGISTICS

####Location & Schedule
The hackathon takes place at Betaworks in the Meatpacking District, NYC (29 Little W 12th St).
Full schedule details are available in the event listing. Key times:

Day 1 (Feb 15): 8:00 AM - 8:00 PM
Day 2 (Feb 16): 8:00 AM - 5:00 PM
Project Submission Deadline: 1:30 PM on Feb 16

####Resources & Support

OpenAI engineers and sponsor representatives will be available throughout the event
Mentoring available throughout Day 1
Technical workshops and support sessions

###CODE OF CONDUCT

We’re committed to maintaining a welcoming, inclusive environment:

Treat all participants with respect
Work with integrity and honesty
Focus on constructive feedback
Help others learn and grow
Follow sponsor platform guidelines
Report any concerns to organizers

###GETTING HELP

Visit sponsor booths for technical guidance
Connect with mentors during scheduled sessions
Reach out to organizers for general questions

Good luck, and let’s push the boundaries of what’s possible with realtime voice AI!