Simone Colucci
Head of Product @ xtream
Milan, Italy
Actions
In love with technology and innovation since university, he has transitioned from software engineering to product management a few years ago. Co-founded xtream, a boutique of digital products and AI, where he helped scale-ups and corporates build better solutions for their businesses.
Awards
Area of Expertise
Topics
Losing the compass: tale of a startup failure
Building digital products is hard. Building innovative products is even harder. You’ve got to balance a unique equilibrium between scarce resources, rush to enter the market, feedbacks from everywhere, hype and discontent.
Both the literature and the community are rich with techniques, frameworks, tips and tricks to navigate the uncertainty. Building with users, listening to feedbacks, iterating with a lean mindset, validating assumptions, de-risking hypothesis. But one thing should remain king: your vision as an entrepreneur and founder.
This is a story of a failure: how we launched a compelling travel+ai product idea on the market that went viral for a bit, but failed to execute and scale. Not because the market did not respond the way we hoped, but because too many iterations and too many feedbacks had obfuscated our vision, leaving us with something we ourselves no longer understood or comprehended.
It's not a pitch against lean building or user feedbacks, but a warning to new founders that clarity of vision is more important than anything else.
Anchoring the Insights: Harbour of Finance Wrap-Up Panel
Join us for the final session of the day in the Harbour of Finance track, where our speakers come together to reflect on the discussions and insights shared throughout the day. This is an opportunity to revisit key themes, delve deeper into emerging topics, and engage in an open dialogue about the future of finance and technology. Don’t miss this interactive wrap-up to connect, ask questions, and share your thoughts with industry experts.
LLM pipelines built for critical performances
When designing LLM-based applications, it’s tempting to go with the best-in-class of any component in the content generation pipeline. Yet, especially in consumer use cases, we must evaluate trade-offs to balance content quality and performances, as they both severely impact UX and one cannot compensate the other.
Picking from our own experience building a consumer app serving users real-time content generated on the fly with a complex LLM pipeline, we will explore lessons learnt, tips and tricks to find the sweet spot between these two contrasting forces.
Pretotyping in the real world - quickly iterating between good and bad startup ideas
90% of new projects and ideas end up failing in the market, with the majority among these failing due to a lack of user need. Resources are being spent building something nobody wants. A few years ago the former Google Innovation Agitator, Alberto Savoia, introduced a technique called “pretotyping”, designed to quickly sample market interest before committing resources to a new idea or project.
In this talk we’ll explore how pretotypes can help startups spot ideas worth pursuing, and specifically how we used it to kickstart an innovative consumer app in the traveltech space. With a combination of techniques like Mechanical Turk and Fake Doors, we managed to gather over a thousand early adopters without a single piece of product.
We’ll talk about lessons learnt, tricks we found particularly effective and caveats we had to learn the hard way.
Actionable user feedbacks for LLM applications
LLM-based use cases are getting introduced and implemented within digital products every day. Yet, revolutionary as these new tools might be, product teams must stay grounded to actual user requirements and measure their effectiveness within the boundaries of product management best practices. This means observing user behaviour, A/B testing to validate product assumptions.
But how can traditional product analytics work with the non deterministic nature of LLM outputs?
In this talk we’ll explore how open-source project Langfuse can help correlating user actions or explicit feedbacks to LLM generations, so that the team can quickly A/B test different versions of prompting and measure what works best for the final users.
Agents, Assemble! GenAI on steroids with AutoGen and multi agents
Quasi ogni prodotto digitale attualmente in sviluppo sta introducendo GenAI sotto qualche forma. Ma quante di queste applicazioni ne sfruttano al 100% le potenzialità?
In questo workshop, il nostro percorso partirà da un'app di esempio realizzata a partire dal gpt-4o o l'o1 di OpenAI, chiamato via api a singolo prompt. Da lì, porteremo il nostro prodotto su nuove vette utilizzando AutoGen e AutoGen Studio.
Questo framework ci aiuterà a rendere passo dopo passo il nostro flusso di generazione più sofisticato, sfruttando agenti LLM specializzati che collaborano tra loro per ottenere risultati di maggiore profondità e precision.
Vedremo come sfruttare AutoGen Studio per prototipare rapidamente questi flussi, e come introdurli nella nostra applicazione disaccoppiandoli dal resto della codebase.
Alla fine del workshop, i partecipanti avranno un'idea precisa del vantaggio nell'utilizzare un paradigma multi-agents, con strumenti pratici per sperimentarlo nei loro prodotti.
Please note that Sessionize is not responsible for the accuracy or validity of the data provided by speakers. If you suspect this profile to be fake or spam, please let us know.
Jump to top