Franco Cerino, PhD

–Engineered multi-objective predictive models using adaptive algorithms, reducing inference time for complex
physics models to∼8ms while maintaining high fidelity.
–Implemented end-to-end AutoML pipelines usingDataRobot(evaluating 80+ algorithms) and Bayesian
hyperparameter optimization withOptuna.
–Lead Architect of the open-source libraryscikit-reducedmodel(scikit-learn compatible). Integrated CI/CD
workflows (GitHub Actions),tox, and achieved93% unit testing coverage.
–Designed an image compression codec (AQMP) that outperformed JPEG efficiency with 4x higher compression
rates in specific domain scenarios.
–Peer-reviewed author of 4 international publications on predictive modeling and applied blockchain.
–Keynote speaker at international scientific conferences.

–DevelopedBlockchainTracerin Python for immutable ML model traceability. Designed hybrid storage
(on-chain/off-chain via IPFS) compliant with industrial Model Cards standards.
–Architected Solidity Smart Contracts (EVM) for Green Hydrogen traceability, including a Proof of Concept (PoC)
for its tokenization (ERC-20).
– Awarded 2nd Place at the Algorand Mega-Ace Hackathonfor developing a blockchain-based green
hydrogen supply chain traceability solution.