Weekly Brunches with Workshops: Amplifying Human Performance in Combinatorial Competitive Programming and How AI Is Reshaping Startup Funding and Building
This event is part of the Weekly Brunches with Workshops series. Read more about the programme here.
Entrepreneurial talk: How AI is reshaping the way Entrepreneurs Build and Raise Capital with Sam Marchant (Inaugural)
AI has fundamentally transformed investment dynamics and, consequently, changed how entrepreneurs build startups. Sam will examine these changes and their implications for startup investors. How can investors work with talented founders when the cost to build companies is falling, but more capital pours into the asset class with the promise of high returns?
Sam Marchant is a former founder now Partner at the AI focused venture builder, Inaugural. Prior to launching Inaugural, Sam was a seed stage VC investor at Salica Investments.
Inaugural is led by the team that built Shazam, iTunes, Google Play and a handful of B2B SaaS unicorns. Using proprietary AI and a focused development process, the Inaugural team ideate, validate and build AI companies across a wide range of sectors and geographies. In the last 12 months, Inaugural have launched 3 AI companies, the largest scaling to over $3.2m in annualised revenue in under 7 months.
Technical talk: Amplifying Human Performance in Combinatorial Competitive Programming with Petar Veličković (Google DeepMind)
Recent years have seen a significant surge in complex AI systems for competitive programming, capable of performing at admirable levels against human competitors. While steady progress has been made, the highest percentiles still remain out of reach for these methods on standard competition platforms such as Codeforces. In this talk, Petar will describe and dive into our recent work, where they focussed on combinatorial competitive programming.
In combinatorial challenges, the target is to find as-good-as-possible solutions to otherwise computationally intractable problems, over specific given inputs. They hypothesise that this scenario offers a unique testbed for human-AI synergy, as human programmers can write a backbone of a heuristic solution, after which AI can be used to optimise the scoring function used by the heuristic. They deploy our approach on previous iterations of Hash Code, a global team programming competition inspired by NP-hard software engineering problems at Google, and they leverage FunSearch to evolve our scoring functions. Our evolved solutions significantly improve the attained scores from their baseline, successfully breaking into the top percentile on all previous Hash Code online qualification rounds, and outperforming the top human teams on several. To the best of our knowledge, this is the first known AI-assisted top-tier result in competitive programming.
Petar Veličković is a Senior Staff Research Scientist at Google DeepMind and Affiliated Lecturer at University of Cambridge (Department of Computer Science and Technology). He is recognised for his pioneering contributions to neural algorithmic reasoning, graph representation learning, and categorical and geometric deep learning. His work has had a profound real-world impact, from improving Google Maps travel-time predictions (featured by Engadget, VentureBeat, CNET, The Verge, and ZDNet) to guiding mathematicians toward new theorems and conjectures (covered by Nature, Science, and Quanta). He also contributed to the first full AI-driven football analytics system and co-developed the first AI-assisted top-percentile result in competitive programming.
The brunches are enabled by:
WHEN: Friday, 7 March 2025
1-2 PM: Brunch in King’s Hall
2-3 PM: Workshops
* Entrepreneurial workshop in Provost's Drawing Room
* Technical workshop in Seminar Room
3 PM onwards: Optional informal discussion in the Turing Room, the E-Lab co-working space
WHERE: King’s College, Cambridge
TICKETS: Workshop registration opens on the Monday before each event. Brunch signups open on the following Tuesday or Wednesday via our WhatsApp group chat.