Fincons Group

My network please consider this gentlemen for a role if available.

I’m happy to share that I’ve obtained a new certification: FinOps Certified Practitioner from The Linux Foundation FinOps Foundation . Thanks to Adèle Brun for the inspiring training! 😊

Seeing our efforts recognized is incredibly rewarding. This achievement reflects the dedication and high-quality expertise that Fincons provides to the Media & Entertainment sector. My sincere thanks to my Fincons colleagues who made this possible: Vito Belviso , Paolo Ferrara , Giulio Spotti , Federico Ricchi , Andrea Corti , Andrea Cimini , Paolo Cardinale , Andrea Fumagalli , Umberto Benelli , Anna Petrarota , Alberto Niero and to our AWS partners who have supported our journey: Kevin Savina , Alberto Falcini , Rui Cardoso, Andrea Nieto Velasco . Unbelievable teamwork!

I’m happy to share that I’ve officially wrapped up my latest research project focusing on the cognitive mechanics of political discourse! I've used Llama-3 and CRUM to move beyond binary sentiment analysis, mapping 'Hot Cognition' and latent cognitive appraisals across 30k+ records of polarized political discourse. Traditional sentiment analysis often hits a "strategic plateau." By shifting the focus from descriptive linguistics to predictive cognitive modeling, this project uncovered: Hot Cognition: How personal goals and social identities drive reasoning. Evaluation vs. Description: Polarization isn't just about different news diets; it's about how we evaluate the same shared reality. LLMs as Annotators: Using Llama-3 as a "cognitive lens" to extract complex variables such as Agency and Goal Congruence at scale. It’s been a fascinating journey into the "cognitive machinery" of the human mind. On to the next challenge! If you're interested, take a look at the Colab Notebook: https://shorturl.at/yuHss #DataScience #AI #NLP #CognitiveScience #PoliticalDiscourse #Llama3

Earlier today I saw a post claiming you can build a "production-grade AI app" over a weekend, on $0 infrastructure, with your first paying customer "sooner than you think." I've been building an AI application for months. With Claude Code, every day. And I can tell you: scaffolding an app in a weekend is absolutely possible. Claude Code is incredible at that. But scaffolding is not production. Production is token cost management so your users don't burn money on every conversation. Production is context optimization so the agent doesn't lose track after 10 turns. Production is error handling, persistence, security, testing. Production is the hundred decisions that don't fit in a LinkedIn post. The tools we have today are amazing. But they don't remove the engineering — they accelerate it. There's a big difference. To everyone posting "build X in a weekend" content: I'd genuinely love to see the follow-up. Show us the app running in production. Share the real numbers, the trade-offs, the things that broke on free tiers. That's the content this community actually needs — not the scaffold, but what comes after. I'm not saying this to discourage anyone from building. Build. But be honest about what it takes. A weekend gives you a prototype. Production takes months. I know because I'm still working on mine. #BuildInPublic #AI #AgenticAI #ClaudeCode #DevTools #SoftwareEngineering