AP

Alex Perritaz

Chief Architect at InFlux Technologies
On the record
Represented by:
Share profile 
Link:
Bio
Edit

Alex Perritaz is an experienced technologist with 15 years of coding expertise and a deep background in IT business management. Specializing in artificial intelligence, Alex's master’s thesis focused on transformers and attention models in 2018/2019, laying the foundation for his career in advanced computing.

Over two years ago, Alex joined InFlux Technologies, quickly rising to the role of Chief Architect. In this capacity, he has spearheaded the design and development of FluxCore and FluxEdge, leading the architecture of a decentralized computation network. Leveraging Kubernetes, Alex's work has enabled handling diverse workloads, including AI, bringing him back to his original passion for artificial intelligence.

Alex’s leadership in building scalable and decentralized infrastructure has been crucial in driving innovation at InFlux Technologies.

Alex is an AI and IT infrastructure expert specializing in compute and sustainability.

  • Specialized vs. Generalized Models: A Balanced Approach in AI
    Alex emphasizes the need for a balanced use of specialized and generalized models, stating, "A selection of specialized models is better than one LLM." He highlights the importance of context in deployment and environmental considerations, noting that specialized models should complement generalized ones for efficiency across sectors like finance, retail, and social media.
  • AI Bias: Understanding and Mitigating Human-Introduced Biases
    Alex explains, "AI bias reflects human-introduced biases from data and labels." Unlike AI hallucinations, bias is systematic. While bias-free AI may be unattainable, strategies like diverse datasets and multiple perspectives can reduce bias. Alex emphasizes the importance of context and balanced outcomes in debiasing efforts.