AI Signals #34: Agent Harnesses and Geometric Thoughts

⌖ London

Hosted by UKAI

Explore how organisations can demonstrate trustworthy AI in practice at UKAI’s free, invite-only report launch in Central London on 29 October 2026. From AI principles to practical trust Taking place from 6pm to 8.30pm GMT in Central London, this VIP drinks reception marks the release of Trustworthy AI: From Principles to Practice, A Framework for Innovation, Accountability and Adoption. The report has been produced by UKAI in partnership with Capgemini and shaped by a roundtable of senior leaders spanning financial services, legal, data and compliance. Rather than focusing only on AI regulation, it examines a wider and increasingly important question: what makes AI trustworthy, and how can organisations demonstrate that trust in practice? Attendees will receive an early look at the report’s findings and hear from contributors involved in its development. The evening will explore the intersection of responsible innovation, accountability and AI adoption, offering useful context for organisations seeking to turn high-level trust principles into credible working practices. What to expect - A first look at findings from the new UKAI and Capgemini report - Insights from contributors who helped shape the work - Drinks and informal networking in a Central London setting - Opportunities to meet UKAI members and professionals advancing the AI trust agenda - Discussion around AI governance, adoption, innovation and accountability This is a reception and report launch rather than a hands-on workshop. Its informal format is designed to support conversation and relationship-building among people dealing with the practical and strategic implications of trustworthy AI. The event is particularly relevant to senior leaders, AI governance and compliance professionals, legal and data specialists, financial services practitioners, innovators, and decision-makers responsible for adopting AI within organisations. It will also appeal to UKAI members and report contributors who want to connect with peers working at the frontier of responsible AI. View the event and registration details

Sponsors: Capgemini

⌖ London · The Royal Institute

Hosted by Cambridge Centre for Data-Driven Discovery

A one-day, in-person workshop in London exploring how frontier AI systems can be secured in real-world environments. Women in AI Security Workshop 2026 takes place at The Royal Institute on Tuesday 3 November 2026, from 9:00am to 5:00pm. Built around the theme “Securing Frontier AI Systems in the Wild”, the workshop will showcase prominent women working across AI security, AI safety and cyber security. It brings together perspectives from government, academia and industry, giving attendees an opportunity to learn about recent technical advances while examining the practical challenges involved in deploying and protecting frontier systems. What to expect The programme combines several formats to support technical learning, discussion and new connections: - Technical talks covering developments in AI and cyber security - Real-world case studies of how frontier AI systems are being secured - Panel sessions featuring experts from a variety of backgrounds - A guest lecture - Short community spotlight talks - A poster session presenting work from across the field - Opportunities to meet researchers and practitioners working at the intersection of AI and security The event forms part of an ongoing effort to build a more inclusive and diverse AI security community. Attendance is open to all genders, roles and levels of seniority, although the content will be particularly relevant to researchers and practitioners whose work connects AI with cyber security. The mix of talks, panels, posters and networking should suit both established experts and those developing their place in the field. Submit a spotlight talk Women working in AI security can apply to take part in the new spotlight talk session, with submissions welcomed at any career stage. Selected speakers will receive up to 15 minutes to present, with no audience questions afterwards. These sessions do not require a finished project. Early ideas, work in progress, interesting findings and open questions are all encouraged, creating a friendly setting in which women can share emerging work with an engaged audience. In support of the workshop’s goal of providing a platform for women in the field, spotlight speaking slots are reserved for women, while general attendance remains open to everyone.… Register for this event

◎ Online

Hosted by Financial Times

Join the Future of AI a live event by Financial Times, held in London from 4-5 November 2026. Register for in-person or digital pass. Meet those at the cutting edge of AI innovation and explore how this transformative technology will redefine business in 2026 and beyond.

⌖ London · Future of AI Conference

Hosted by Future of AI Conference

Future of AI Summit 2026 Day 2 brings a full day of AI technology and innovation to Convene 22 at 22 Bishopsgate in London. Taking place on Thursday, 5 November 2026, the conference runs from 8:00am to 5:30pm GMT. It is aimed at people interested in the latest developments shaping AI, from technology professionals and business leaders to founders and other innovation-focused attendees. What to expect This is the second day of the Future of AI Summit 2026, presented in an in-person conference format. The published listing categorises it as a tech event but does not yet provide a detailed agenda, session topics or speaker line-up. The venue is Convene 22, 22 Bishopsgate, London, England. The listing also advertises 20% off tickets when booking with the promo code DISCO through the linked ticket offer. Register and view ticket options.

⌖ London · QMUL, ArtsOne, Bloc, Mile End Road, London E1 4PA

Hosted by A Season of Bangla Drama 2026 (LB Tower Hamlets)

Explore how AI could reshape electronic and print journalism at this 90-minute, bilingual talk for Bangladeshi media professionals in London. Presented by the London Bangla Press Club, the session examines how artificial intelligence may affect everyday newsroom work across newspapers, online news portals and television. Topics include: - Research, data collection and analysis of large archives - Fact-checking and cross-checking statistics - Detecting manipulated images and inconsistencies - Tracking public trends and monitoring social media for breaking news - Live speech-to-text transcription and instant translation - Real-time data updates for electronic media - Analysing financial records and public data for investigative journalism The in-person talk is aimed at reporters, journalists, editors, producers and other media professionals, particularly London Bangla Press Club members working with Bangladeshi media. It will be delivered in English and Bangla and is open to attendees aged 16+. No individual speaker line-up is published. Event details - Date: Tuesday 10 November 2026 - Doors: 6:00pm - Talk: 6:30–8:00pm - Venue: QMUL, ArtsOne, Bloc, Mile End Road, London E1 4PA - Format: In-person, 90-minute talk The London Bangla Press Club has supported journalists of Bangladeshi heritage through training, networking, professional development and cultural activities for more than 33 years. The event is presented by A Season of Bangla Drama 2026 (LB Tower Hamlets). Get tickets on Eventbrite

⌖ San Francisco · Hyatt Regency Burlingame

Hosted by Open Data Science Conference (ODSC)

Learn to build production-ready AI systems in just seven weeks with the AI Engineering Accelerator bootcamp, a hands-on training program that bridges the gap between using AI and building it. Running from June 9 through October 29, 2026, this intensive program combines virtual instruction with a multi-day in-person capstone at the Hyatt Regency Burlingame during ODSC AI West 2026. Designed for professionals ready to move beyond tutorials, the Accelerator provides comprehensive training in the core disciplines of modern AI engineering: large language models (LLMs), agentic AI systems, and production deployment. You'll gain practical experience with the latest tools and frameworks used by leading AI teams, learning not just how to use AI, but how to architect, build, and ship AI-powered applications. Program Structure The bootcamp follows a hybrid learning model that maximizes flexibility while ensuring hands-on engagement. Starting with the Summer Cohort on June 9, you'll participate in virtual training sessions led by experienced AI practitioners. The program culminates in an intensive in-person segment from October 27-29, 2026 at the Hyatt Regency in Burlingame, California, integrated with the main ODSC AI West conference. This dual-format approach allows you to build foundational knowledge remotely while maintaining your current work commitments, then apply those skills in collaborative, hands-on projects during the in-person sessions. You'll also gain full access to ODSC AI West's conference tracks, expo, and networking events. What You'll Learn The curriculum focuses on production-grade AI engineering, covering: - LLM development and deployment — Fine-tuning, prompt engineering, and optimization techniques - Agentic AI systems — Building autonomous agents and workflow automation - AI Engineering fundamentals — MLOps, model serving, monitoring, and scaling - Real-world implementation — Moving from prototype to production with confidence Organized by Open Data Science Conference, which has spent 11 years training thousands of AI engineers and data scientists, the Accelerator draws on ODSC's deep expertise in practical AI education. The program features specialization tracks tailored to different AI engineering domains, allowing you to focus on the areas most relevant to your career goals. Who Should Attend This bootcamp is designed for professionals who already work with AI tools and are rea