"Team Eagle team on AMD Developer Hackathon: ACT III Hackathon"

Team Idea

Building a hardware-optimized backend pipeline running on AMD Instinct GPUs using the Evolus framework. Deploying open-weights LLMs as Autonomous Cryptanalysts to analyze structural linear/differential weaknesses and optimize S-box configurations in lightweight ciphers (like Ascon) used to secure fragile medical IoT telemetry. This is a solo project executing end-to-end systems programming, low-level data extraction, and ROCm platform optimization to visualize mathematical token uncertainty.

Team Members

Diwakar Singh

diwakarsharmx

Submission

Ascon-Guard AI: Hardware Security for Medical IoT

Autonomous hardware security pipeline pairing NIST SP 800-232 Ascon-128 edge encryption with uncertainty-calibrated neural cryptanalysis on AMD Instinct MI300X accelerators and Evolus Multi-Agent orchestration to protect next-gen connected medical IoT.