Browse applications built on Google Gemma 2 technology. Explore PoC and MVP applications created by our community and discover innovative use cases for Google Gemma 2 technology.
An agentic pipeline that takes any CUDA repo and autonomously ports it to ROCm/HIP — analyze → transpile → compile → test → benchmark on real AMD GPUs — with an LLM repair loop fixing everything the mechanical tools can't.
In 2024, ~2.4 million people who developed TB went uncounted. Jaga is a phone-first triage prototype: one guided cough plus clinical inputs becomes a follow-up urgency estimate—no lab or X-ray—so workers prioritize who to test first.
Real-time quantum device digital twin with AMD MI300X GPU acceleration and intelligent triage agent that maintains state sync under extreme throughput conditions.
We built ReCodeX, the Legacy Code Modernization Engine. It's an AI platform that automatically transforms legacy "spaghetti" code into clean, secure, and cloud-ready infrastructure. Powered by AMD high-performance cloud hardware.
SimuFactory is a secure, on-premise industrial trainer powered by AMD ROCm. Using a Llama-3/Gemma dual-model architecture, it provides Socratic coaching and safety auditing via a live Digital Twin to train technicians without risking real hardware.
Automated your PCB Design workflow with CircuitBlocks!!
A neural network that grows its own brain mid-training. When loss spikes, a local Gemma-2-2b-it LLM writes, validates, and hot-swaps in a new expert module — no human, no restart. Verified on real AMD ROCm hardware.
Secure Scope AI is an AI-powered security incident management platform that helps organizations detect, analyze, prioritize, and respond to cyber threats faster with intelligent automation, centralized workflows, and actionable AI insights.
An AI platform that creates video parodies by re-voicing existing footage or generating 9:16 vertical short-form videos from scratch using Moonshot Kimi K2.6, FLUX.1, and Google Gemma models with an asynchronous multi-threaded Streamlit interface.
GlazeSmith is a physics-grounded digital twin for ceramic glazes. It predicts how a recipe fires - crazing risk, surface, color - then renders the result live on a rotatable 3D vase, verified end-to-end. Classifiers trained on AMD ROCm.
RasoSynthTune is an autonomous multi-agent data refinery. It discovers extracts filters and constructs fine-tuning datasets, running end-to-end verifications and native model distillation on AMD Instinct GPUs via Fireworks AI
TURF is a native iOS & Android AI fitness app where your runs conquer real city territory. Powered by Fireworks AI on AMD hardware for live coaching, route intelligence & real-time 3D map territory capture.
Upload any PDF and chat with it using Gemma 4. Get accurate, cited answers, instant summaries, and auto-extracted action items — built on Supabase and Google's Gemini API.
AI negotiation infrastructure that transforms messy freelance conversations into trusted contracts through multi-agent reasoning, deterministic guardrails, and private Gemma inference on AMD GPUs.
An agent-driven AI system that automates model training, reasoning, and deployment decisions.
GuideMate is an AI-powered learning mentor that creates personalized roadmaps based on each learner's goals, learning style, and available time. It delivers daily lessons, practice, and AI guidance to help learners achieve their career goals.
Kernel Olympics is a multi-agent AI pipeline that automatically ports NVIDIA CUDA kernels to AMD ROCm/HIP — transforming weeks of manual engineering into ~2 minutes per kernel with 3-stage validation gates and real GPU verification.
Cloakwell AI is a local-first DLP (Data Loss Prevention) platform that intercepts AI prompts, classifies PII (Personally Identifiable Information) and sensitive data in real-time, by routing queries to private AMD GPUs running Gemma 2.
Agentic Microcontroller Tester Agentic Microcontroller Tester
Longhand helps Americans in the UK file their US taxes without the guesswork. Python does the maths, the AI only explains, and every number links to the exact line of law behind it. It compares FEIE vs FTC, flags your forms, and fills the real IRS PDFs.
NPCVerse AI creates intelligent, memory-driven game characters with evolving personalities, emotions, and adaptive dialogue. Using multi-agent AI, NPCs remember player interactions, build relationships, and deliver dynamic, lifelike gameplay experiences.
An LLM proposes, an exact solver decides, an independent verifier certifies — business decisions you can prove feasible and optimal, not just plausible answers from a chatbot.
Not just a period tracker — Elara connects your cycle to your emotions, so support actually feels personal
hrome Manifest V3 extension that shortens AI prompts on ChatGPT, Claude.ai, and Gemini — local regex fluff removal plus optional semantic compression via a containerized companion (Fireworks Gemma) or local Ollama.
GemmaRoute is a 4-layer hybrid AI router that cuts LLM API costs by 80%. Local Gemma 2B on AMD ROCm handles trivial and simple queries for free. A Gemma quality judge auto-escalates to cloud only when response quality demands it.
AcademicForge is an AI-powered academic writing assistant that helps students and researchers plan, draft, refine, and organize high-quality scholarly work faster.
An AI-powered Resume Analyzer that evaluates resumes, matches them with job descriptions, and generates personalized interview questions and improvement suggestions using a local LLM.
LedgerBuddy AI empowers small businesses with an intelligent smart ledger. It automates invoice data entry using AI document extraction and features a multilingual financial Copilot.
ForgeMind AI is an AI-powered software architecture assistant that transforms project ideas into complete technical blueprints, including system architecture, security recommendations, development roadmaps, and implementation guidance and etc.
This is just a test submission This is just a test submission