Dual Pitcher AI Agent

Created by team NPA on July 06, 2026

AI/ML API AMD Developer Cloud Claude Code Codex DeepSeek V3 Gemma gpt4all Llama 4

Hybrid Token-Efficient Routing Agent

Utilizing new frontier AI models comes at a cost... quite literally. Companies like Uber, Meta, and Microsoft have expended their AI resource budget due to internal over-use of AI.Thus, there is a growing need for enterprises to control AI spend without sacrificing user experience: not every task needs a premium proprietary model. This project aims to build that smart router by running variety of local models and making as few external Fireworks API calls as possible while still clearing the accuracy gate. Our agent will be tested across all eight categories: Factual knowledge, Mathematical reasoning, Sentiment classification, Text summarization, Named entity recognition, Code debugging, Logical reasoning, and finally Code generation.

Category tags:

Optimizing Resource Allocation, Open-ended Agent Protocol

Github Presentation Demo

Explore more applications

Raccoon Vision Translator

Watches a clip and writes four captions in four different voices: formal, sarcastic, tech humor, casual humor. All four come from the same fact-checked scene description, using Gemma 4 through Hugging Face.

Ryzen from the Dumpsters

HuggingFace HubGemmaWhisperrest api

WhisperLens

Containerized video captioning agent with AMD ROCm PyTorch support.

Vision Scions

GemmaGemini AIWhisperAMD ROCm

Beroute

Python-heavy project built for experimentation and rapid prototyping, centered on high-performance computing. While most of the codebase is in Python, suggesting parts of the system are optimized for speed or lower-level hardware interaction.

Berking

AntigravityStreamlit

Token Frontier Model

The measurement instrument for a token-vs-accuracy cascade router.

Paripakv

AMD Developer CloudAnthropic ClaudeGemma

Prueba

sdsddwdw dq dwdwqd dqdqwd qwd qwd qwd qwd wdqqwdqdwq wdqwdqwdqwdqwdqwdqwdqwdq wdq dw qdw qwd qw dq wd qwd w

R-bot

AMD Developer Cloud

Upcoming AI Hackathons
For Innovators & Creators

Explore more applications

Raccoon Vision Translator

Watches a clip and writes four captions in four different voices: formal, sarcastic, tech humor, casual humor. All four come from the same fact-checked scene description, using Gemma 4 through Hugging Face.

Ryzen from the Dumpsters

HuggingFace HubGemmaWhisperrest api

WhisperLens

Containerized video captioning agent with AMD ROCm PyTorch support.

Vision Scions

GemmaGemini AIWhisperAMD ROCm

Beroute

Python-heavy project built for experimentation and rapid prototyping, centered on high-performance computing. While most of the codebase is in Python, suggesting parts of the system are optimized for speed or lower-level hardware interaction.

Berking

AntigravityStreamlit

Token Frontier Model

The measurement instrument for a token-vs-accuracy cascade router.

Paripakv

AMD Developer CloudAnthropic ClaudeGemma

Prueba

sdsddwdw dq dwdwqd dqdqwd qwd qwd qwd qwd wdqqwdqdwq wdqwdqwdqwdqwdqwdqwdqwdq wdq dw qdw qwd qw dq wd qwd w

R-bot

AMD Developer Cloud