Vishal Vishnu Kagade@vishalkagade31674

Brainstormer is a voice agent that knows which part of your life you are in, remembers it, and gets better by itself. Try the demo. Username: anything. Password: Vishpatya@123. Limited to 15 users. Contact: [email protected] Why we built it In a text chat, you open a new chat for each topic. On a voice call there is only one thread. You jump from topic to topic, and the agent has to keep up. Most voice agents use one prompt for everything. We saw three problems: - One style does not fit every topic. At the gym you want a short answer. For a technical question you want the agent to wait and explain. - It forgets. The last call is gone when you hang up. - Mistakes stay. When a reply goes wrong, nobody fixes the prompt. What Brainstormer does It has a different setup for each part of your life. We call each setup a mode: Gym, Deep work, Shopping. Each mode has its own prompt, its own waiting time, its own tools and its own memory. - You talk about a new subject, and the mode changes by itself. The call does not restart. - You talk about something new, and a new mode is created during the call. - After each call, every reply is checked. If a mode keeps making the same mistake, a better version of it is written. - Every version is saved. You can see what changed and why, and go back to an older one. - A mode you do not use for 30 days goes to sleep. It wakes up when you talk about it again. How it is built AssemblyAI Voice Agent API runs the call. The Jev decision model picks the mode. A model on Fireworks AI writes the better versions. Exa does the web search.

Voice cloning is now so realistic that just three seconds of audio can imitate someone’s voice, and AI voice scams increased by 1,210% in 2025. Scam Call Shield protects users during live calls in two ways. Whisper transcribes speech locally, while an open LLM running on Fireworks analyzes the conversation for scam patterns and warns the user. At the same time, a wav2vec2 model, fine-tuned and evaluated on AMD GPUs with ROCm, checks every three seconds for AI-generated speech. This catches cloned voices that sound normal. Our model achieved 98.8% accuracy on real-world deepfakes and 87.4% on unseen speakers.

🌎 Hybrid Hackathon | 💻 Online Build: September 10–16, 2026 | 📍 On-site: September 15–17, 2026 The official hackathon of AI Infra Summit — co-organized by Kisaco Research and lablab.ai. Build production-grade AI applications on the modern AI infrastructure stack, alongside 8,000+ attendees, researchers, founders, and infra leaders. 📍 On-site Venue (September 15–17): Santa Clara Convention Center 5001 Great America Pkwy, Santa Clara, CA 📍 On-site participation is by invitation only. Travel and accommodation expenses will not be covered.

🌎 Online Hackathon | 💻 Month-long challenge | Dates: Sep 1–30, 2026 | 🏆 $10,000 Prize Pool ($5k cash + $5k in AAI credits) A month-long online challenge run by lablab.ai together with AssemblyAI, the Voice AI infrastructure company for builders. Every participant builds on AssemblyAI. You can join at any point — registration stays open for the whole build window, so you can start on day one or pick the project up halfway through and still submit. 🌍 Fully online hackathon. Join and build from anywhere in the world.

To participate in the hackathon, simply click Sign up with AMD. If you are not already a member of the AMD AI Developer Program (ADP), you will need to create an account. ADP members can access AI tutorials, experts, and community support. In addition, new members can claim a free month of DeepLearning.ai pro, $100 in AMD GPU cloud credits, and $50 in Fireworks AI API credits. All hackathon participants will receive on hackathon launch additional compute and API credits for their projects. Ready to build what's next on AMD? Sign up with AMD below to secure your spot.