Jakub Nowacki, CTO and Co-founder of Nusmark, builds an agentic AI assistant that transforms personal productivity by combining multiple specialised components into a seamless 24/7 digital secretary accessible in whatsapp. Starting his coding journey at nine years old creating video games and building AI opponents when friends chose basketball over programming, Jakub mentored over 100 students at university, won the AI Summit London twice (first person to achieve this), and left corporate aerospace after discovering the bureaucratic inefficiencies of requiring forms to send emails. He explains agents as infrastructure around LLM "brains" that provide tools and execute actions in loops, using Gemini for its massive context window and Google TPU speed advantages while building modular systems that can swap models instantly based on performance benchmarks. Jakub reveals Nusmark's sophisticated architecture featuring custom thinking layers, parallel/sequential processing for reduced latency, and RAG-based tool selection to handle future MCP server ecosystems, while implementing both short-term memory (stored in context) and long-term memory (extracted patterns from months of conversations). Their breakthrough phone calling feature uses speech-to-text, LLM processing, and text-to-speech components to book NHS appointments and restaurant reservations, with users questioning whether real humans are involved, while building toward agent-to-agent token-based communication that bypasses human language inefficiencies. Jakub positions Nusmark as the "super connector" that integrates across competing platforms (Meta, Google, Apple) where individual companies cannot due to competitive barriers, offering proactive capabilities like following up on delayed callbacks and scheduling meetings between users' agents that negotiate optimal times based on preferences like avoiding dinner conflicts and maintaining buffer zones around internal meetings, ultimately creating a mobile-first calendar experience designed specifically for touch interfaces rather than desktop-to-mobile ports.

Harvey Pitt, CTO and Co-founder of Alive Industries, pioneers the "zero-waste startup" approach by building a bootstrapped AI consultancy that gets paid to discover product opportunities while solving real customer problems, positioning AI as a "last resort" technology that should only be used when deterministic solutions fail. Having started at Oxford after witnessing early GPT-3 capabilities, he explains their philosophy of taking existing broken processes and making them faster rather than creating entirely new workflows, emphasizing that good AI engineers tell clients they don't need AI since it's inherently probabilistic technology where "you're rolling the dice" compared to deterministic systems that guarantee outcomes. Harvey reveals two fascinating projects: using pre-2020 computer vision technology to reconstruct 3D sail models from 2D yacht images for real-time performance validation (processing entirely on-device for ocean environments), and building LLM-powered compliance systems for waste management brokerages that connect large companies with local disposal services to reduce carbon emissions by extracting structured data from complex regulatory documents. Through their Machine Talents community (machinetalents.com), Alive Industries hosts hackathons at venues like Google with nearly £10,000 in prizes, creating a talent pipeline that identifies 10x engineers early before market rates reach "billion pound offers" while giving back to the ecosystem and proving product viability to potential clients within 24 hours. Harvey positions hackathons as essential infrastructure for Europe's startup ecosystem, providing networking opportunities for developers who may lack traditional social skills, enabling rapid technological iteration cycles, and offering startups cost-effective talent acquisition alternatives to expensive referral bonuses, while creating transparent environments where engineering capabilities and collaboration skills are immediately visible through live GitHub commits and real-time problem-solving demonstrations.

Dana Aleff, CEO and Founder of Circonomit, builds the "ethical European Palantir" by creating enterprise software that maps reality into interconnected digital objects, enabling companies to understand causational relationships across their entire business ecosystem rather than relying on fragmented Excel spreadsheets and reactive monthly financial statements. As a mechanical engineer turned obsessive systems thinker who's conducted over 2,000 customer conversations in two years, she reveals Circonomit's ontology layer that transforms any physical object—from bottles to buildings to people—into connected digital entities with properties, dependencies, and calculations, creating what she calls "encoding reality that we forgot to write down while building our complex world." Dana explains their Figma-like visual interface that allows non-technical users to build sophisticated business models by dragging and connecting objects on a canvas, making complex optimization problems as accessible as games while unlocking 10,000x faster decision-making compared to traditional systems where "it takes longer to look up information than to make the actual decision." Driven by environmental concerns and belief that financial, social, and sustainability factors are inherently interconnected, she positions Circonomit as democratizing optimization problems by giving decision-makers tools to understand trade-offs rather than making gut-feeling choices based on incomplete information. Her ultimate vision extends beyond enterprises to "building nations" by applying the same causational mapping principles to societal challenges, believing that both companies and countries are nested systems requiring quantified understanding of dependencies rather than black-and-white political debates, while acknowledging the ongoing challenge of encoding human emotional and irrational behaviors into their systematic approach that currently combines data patterns, human knowledge, and AI-assisted thinking to create comprehensive "strategic twins" of organizations.

Callum Woodcock, CEO and Founder of Wineify, democratises fine wine investing through a multi-award-winning platform that makes investment-grade wine accessible from £3,000 using innovative syndicate structures, positioning wine as a legitimate portfolio diversifier with equity-like returns and capital gains tax exemptions. Operating from rural France, he explains Winefi's UK Bear Trust nominee structure that allows syndicate members to directly own wine rather than the platform owning it, preserving CGT benefits while enabling bulk purchasing power through pre-raised funds that secure discounts from merchants and private sellers seeking liquidity. Callum reveals their "Moneyball for wine" approach, leveraging 25 years of real-time trading data across 100,000 wines and 18 million data points to build predictive models that identify outperforming investments, incorporating critic scores weighted by regional influence (Robert Parker for Bordeaux, specialists for Tuscany/Piedmont) and developing their proprietary Wine Investment Score (WIS) rating system from 0-100. He challenges the myth that Gen Z drinks less alcohol, citing data showing they spend higher percentages of income on alcohol than previous generations, while predicting taste evolution from heavy Bordeaux toward lighter wines and emerging regions like South Africa, where producers like Sadi are seeing 3-5x price flips overnight despite winemaker resistance. Having grown 100% year-on-year approaching their second anniversary with £1.5 million in funding, Callum outlines expansion plans including multi-currency support, integration with platforms like Hargreaves Lansdown and Revolut, and their vision to become the category-defining brand for wine investment—"the Vanguard of fine wine"—while maintaining sustainable growth through LinkedIn content marketing that builds authentic relationships with clients who become shareholders, demonstrating validation through their CrowdCube funding success.

Ugne from ByFounders show us how Nordic startups evolved from overlooked markets to global unicorn factories, with Stockholm rivaling Silicon Valley in unicorns per capita through a fundamental shift where successful founders like those from Spotify and Klarna stay home to build ecosystems rather than emigrating, creating a virtuous cycle of mentorship and capital. Operating from their pre-seed fund that backs Nordic and Baltic founders globally before anyone else sees potential, she reveals three seismic transformations: the corporate exodus creating experienced technical founders who treat startups as legitimate career paths not fallback options, European risk aversion morphing into thoughtful company building that attracts patient capital, and the emergence of specialized funds like By Founders where every team member from partners to finance must have operational startup experience. She dissects why hackathons trump conferences for sourcing ("watching someone problem-solve exhausted at 3 AM reveals more than any MBA"), how their 40+ member collective of unicorn founders provides unfair advantages in pattern recognition, and why Europe's application layer strength positions it to leapfrog US dominance in specific AI verticals. Ugne traces this evolution from nobody considering Swedish tech a decade ago to today's reality, emphasizing how generational wealth creation happens when ecosystems mature enough that founders build where they live rather than where they're told to go, positioning By Founders as the catalyst connecting quirky, obsessive builders with the capital and mentorship needed to transform regional talent into global champions.

Aaron Jones, Good Omen & Project Lovable Winner, transforms a spontaneous hackathon meeting into a €40,000 breakthrough by building video editing technology and AI journalism platforms in just 20 hours.

Meeting teammates Gio (robotics and chip manufacturing expert) and Marco (described as "the smartest hacker I've ever met") at a Dawn Capital hackathon matching evening.

The team won their first £2,000 prize before receiving a guaranteed invitation to Project Lovable's "unreasonable" themed competition in Stockholm on three days' notice. He reveals their technical innovation using the LTX video framework to reduce video editing inference time by 4x—cutting standard 12-20 minute processing for five seconds of video down to practical speeds by generating four frames instead of eight while looking only at the last frame, then applying this breakthrough to keyframes and dynamic image editing.

The team submitted two complete projects: an AI journalism platform featuring avatars that conduct interviews (addressing industry challenges from TechCrunch leaving Europe to Business Insider downsizing), and their advanced video editing system capable of real-time modifications like changing shirt colors during live streams.

Aaron emphasizes their winning combination of deep research knowledge, technical implementation skills, and consumer taste rare qualities that enabled them to not only win the €20,000 Project Lovable prize but simultaneously earn €20,000 from the journalism company commissioning their platform, achieving a 20x return from their initial hackathon victory while positioning Good Omen for imminent public product launches.

We Max (Investor at wave Ventures) on VENTURES to break down the Lovable recap current state of Stockholm's startup ecosystem and investment challenges. "The talent density is, I would say, one of the highest in the world in Stockholm right now. And incredibly underrated, which works in your favour." "I think the problem is now that we see a lot of international investors coming into Sweden to do deals, which might not benefit the Swedish ecosystem in the long run." "Right now, I don't see that investors in Sweden and in the Nordics can facilitate the need for capital that these companies have." You can listen to the full conversation with Jonas on today's show. Follow viraj & VENTURES on LinkedIn for more.

TODAY ON VENTURES 🚀: 9 Guests, 5 hours, 1 don't miss streamThe docket is locked, loaded & it's special! 5 hours of streaming, bringing you the latest on what is happening in the landscape of technology across Europe. From Alternative Assets to Project Europe's Project LoveableWhere else are you going to find this 🤷‍♂️ 12:00: Callum Woodcock - CEO, WineFi 🍷12:30: Harvey Pitt - CTO/Co-founder, Alive Industries13:00: Sasha Cayward MSc. - Org & Founder Psychologist, Evio13:30: Jakub Nowacki, CTO/Co-founder, Nusmark14:00: Dana Aleff - CEO/Founder, Circonomit 14:30: Ugne Musneckyte - VC Investor, byFounders15:00: 📹 Aaron Jones - Good Omen, Project Europe's Project Loveable 15:30: Luke Nelmes - CMO, Bigged16:00: Max Larsson - Investor, Wave Ventures; Ecosystem builder, Founders HouseJump in the chatFeed us your thoughts & questions before & during- It's LIVE. Follow us on YouTube & LinkedIn to watch DM - if you want to appear on the show or sponsor it LFG 🚀 🚀

David, Founder and CEO of Car Crowd, revolutionises collectable car investing by democratizing an asset class previously exclusive to billionaires like the Sultan of Brunei and celebrities. Having managed collections for over 25 years since childhood, he reveals how Car Crowd's syndicated platform eliminates traditional barriers through expert sourcing, storage, and maintenance while delivering 15.8% annual returns in 2024. He breaks down their evolution from triple-taxed SPV structures to bare trust models that eliminate capital gains tax, their data-driven investment thesis focusing on scarcity and the 25-year import rule, and their global sourcing network that secures museum-quality examples through paint depth analysis and comprehensive inspections. David outlines their ambitious vision to manage 1,000+ assets worth £1 billion by 2028, expanding into watches and wine while building toward a "stock exchange for passion assets" where investors could trade Ferrari shares for Rolex allocations using AI-powered real-time valuations. He emphasizes their community-driven approach where syndicate members vote on sale timing, attend regular events at their Nottinghamshire showroom, and benefit from transparent liquidity management that's achieved a 0.5% exit request rate across nearly 1,000 active investors, positioning Car Crowd as the bridge between traditional asset management and the emerging passion economy.

Martin, co-founder and CEO of Bench, is building "Cursor for hardware engineering" to bring AI assistance to rocket builders, roboticists, and physical product developers. Connecting from San Francisco during a Valley exploration trip, he discusses raising a £1M+ seed round led by Superseed while evaluating whether to split operations between London and US markets. Having founded his university's hyperloop team before meeting Imperial rocket-building co-founder Rayhan, he explains why hardware development has lagged behind software in AI adoption and his vision for automating complex engineering workflows across multiple file types and simulation platforms.