Agentic AI is moving fast from prototypes into production — and with it comes a new set of technical, organizational, and regulatory challenges. This edition of Thursday Gathering brings together founders, engineers, researchers, and operators from Poland’s and Europe’s AI ecosystem to explore what it really takes to build, ship, and scale autonomous systems in the real world.
From trustworthy AI and EU AI Act considerations to production failures, governance gaps, and system design trade-offs, this evening focuses on what actually happens when AI leaves the lab and enters complex environments. Expect honest conversations, practical insights, and no polished hype — just real lessons from the frontlines of building agentic systems.
In the Demo Table zone you will discover various products, technologies and projects being presented by the founders or startup-representatives.
In the Info Table zone you will find out about projects, organisations and ideas.
Want to present your organisation? Send us an application!
However remember, this is NO HARD SALES zone.
Unwind, connect, and vibe to the rhythm! This lively session blends casual networking with great music, offering the perfect atmosphere to meet fellow innovators, exchange ideas, and forge new collaborations. Grab a drink, enjoy the beats, and build meaningful connections in a relaxed, energetic setting.
These 1:1 consultations are for founders, product teams, and operators who are designing, testing, or shipping AI features and want practical clarity on what comes next. In each session, I’d help attendees quickly assess an AI use case, spot likely red flags or grey areas, and turn that into actionable next steps.
The sessions would be especially useful for people who want to understand questions such as:
- whether their AI use case may raise EU AI Act or broader trustworthy AI concerns
- what they should clarify before committing more product or engineering effort
- how to move from uncertainty to a practical action plan with lightweight evidence
This panel brings together practitioners who have taken agentic AI past the demo stage and into the messy reality of legacy systems systems, fragmented data landscapes, and organizations that were never designed to move fast. Panelists will share the friction they didn't anticipate (governance gaps and data silos, change management and the politics of deploying autonomous systems in environments where mistakes have physical consequences and more). If you're trying to understand what actually separates a successful rollout from a stalled pilot, this is the conversation to be in.
This panel goes behind the scenes of customer-facing agentic deployments in banking, e-commerce, and consumer platforms: how do you govern an agent that acts on behalf of a user? What does auditability look like when decisions happen in milliseconds? How do you handle failure modes when the stakes aren't a bad search result but a missed payment or a disputed transaction? Panelists will address the design, deployment, and governance realities that don't make it into the whitepapers.
For years, the AI conversation centred on models: how capable they are, how fast they improve, how much they cost. That conversation is now incomplete. It has shifted from model quality to system design: how you coordinate multiple agents with distinct roles, how you manage state across long-running workflows, how you maintain control when the environment is non-deterministic by nature. This talk draws on real production deployments to map the new hard problems and to frame what it genuinely takes to move from a working prototype to a system that holds up under real-world load.
This one is for the people building the systems, not presenting them. Engineers talking honestly about what broke in production, why they chose the frameworks they did, and what they wish someone had told them before they shipped. Model selection, orchestration trade-offs, context management failures, cost explosions, the moment they realized their evaluation strategy was wrong... it's all on the table. If you've ever sat through a polished case study and thought "that's not what actually happened," this panel is the antidote.
Agentic AI is moving fast from prototypes into production — and with it comes a new set of technical, organizational, and regulatory challenges. This edition of Thursday Gathering brings together founders, engineers, researchers, and operators from Poland’s and Europe’s AI ecosystem to explore what it really takes to build, ship, and scale autonomous systems in the real world.
From trustworthy AI and EU AI Act considerations to production failures, governance gaps, and system design trade-offs, this evening focuses on what actually happens when AI leaves the lab and enters complex environments. Expect honest conversations, practical insights, and no polished hype — just real lessons from the frontlines of building agentic systems.
In the Demo Table zone you will discover various products, technologies and projects being presented by the founders or startup-representatives.
In the Info Table zone you will find out about projects, organisations and ideas.
Want to present your organisation? Send us an application!
However remember, this is NO HARD SALES zone.
Unwind, connect, and vibe to the rhythm! This lively session blends casual networking with great music, offering the perfect atmosphere to meet fellow innovators, exchange ideas, and forge new collaborations. Grab a drink, enjoy the beats, and build meaningful connections in a relaxed, energetic setting.
These 1:1 consultations are for founders, product teams, and operators who are designing, testing, or shipping AI features and want practical clarity on what comes next. In each session, I’d help attendees quickly assess an AI use case, spot likely red flags or grey areas, and turn that into actionable next steps.
The sessions would be especially useful for people who want to understand questions such as:
- whether their AI use case may raise EU AI Act or broader trustworthy AI concerns
- what they should clarify before committing more product or engineering effort
- how to move from uncertainty to a practical action plan with lightweight evidence
This panel brings together practitioners who have taken agentic AI past the demo stage and into the messy reality of legacy systems systems, fragmented data landscapes, and organizations that were never designed to move fast. Panelists will share the friction they didn't anticipate (governance gaps and data silos, change management and the politics of deploying autonomous systems in environments where mistakes have physical consequences and more). If you're trying to understand what actually separates a successful rollout from a stalled pilot, this is the conversation to be in.
This panel goes behind the scenes of customer-facing agentic deployments in banking, e-commerce, and consumer platforms: how do you govern an agent that acts on behalf of a user? What does auditability look like when decisions happen in milliseconds? How do you handle failure modes when the stakes aren't a bad search result but a missed payment or a disputed transaction? Panelists will address the design, deployment, and governance realities that don't make it into the whitepapers.
For years, the AI conversation centred on models: how capable they are, how fast they improve, how much they cost. That conversation is now incomplete. It has shifted from model quality to system design: how you coordinate multiple agents with distinct roles, how you manage state across long-running workflows, how you maintain control when the environment is non-deterministic by nature. This talk draws on real production deployments to map the new hard problems and to frame what it genuinely takes to move from a working prototype to a system that holds up under real-world load.
This one is for the people building the systems, not presenting them. Engineers talking honestly about what broke in production, why they chose the frameworks they did, and what they wish someone had told them before they shipped. Model selection, orchestration trade-offs, context management failures, cost explosions, the moment they realized their evaluation strategy was wrong... it's all on the table. If you've ever sat through a polished case study and thought "that's not what actually happened," this panel is the antidote.