Applications of Computing in Industry

Tuesday 13th January

Time: 1.00pm Location: HUX 308
Title: From Kotlin to CodeSpeak: programming languages and AI coding tools
Speakers:  Andrey Breslav

Andrey Breslav, the creator of Kotlin, will talk about the evolution of programming languages, what makes them success and fail, and how LLMs will affect the languages of the future. Some notable topics we’ll cover: (1) how to design a good language for a given audience, (2) which common pitfalls are tolerable, and which ones are deadly for a programming language, (3) what will programming look like in 10 years given the current AI boom.

Andrey created the Kotlin programming language and led its development at JetBrains from 0 to millions of users. He’s the founder of CodeSpeak.dev, a next-level programming language powered by LLMs.

Tuesday 20th January

Time: 1.00pm Location: HUX 308
Title: Optimising AI Systems: searching for a needle in a haystack during a hurricane!
Speakers:  Dr Anton Lokhmotov 

Artificial Intelligence (AI) is making all the headlines today. Most people tend to think of AI as of huge neural networks such as Large Language Models (LLMs). However, such neural networks ("neuralware") also need specialised chips such as Graphics Processing Units (GPUs) or Tensor Processing Units (TPUs), as well as other hardware. In-between neuralware and hardware, sits "middleware": frameworks, runtimes, compilers, math libraries, etc.
 
Deploying AI (i.e. configuring a complete neuralware/middleware/hardware stack) is hard, as it requires deep expertise across the entire stack. Optimising AI (i.e. searching for fast, efficient and cost-effective deployment configurations) is even harder, as this stack is constantly changing. Searching for optimal configurations can be compared to searching for a needle in a haystack during a hurricane!
 
At KRAI, we are developing KISS (KRAI Inference Serving Solution), an automated platform for AI inference optimisation. KISS is built on the core principles of reproducibility, experiment automation and intelligent design space exploration. Automatically optimising AI inference across the entire stack can bring an order of magnitude improvements in inference costs, energy/water consumption and user satisfaction.
 
Dr Anton Lokhmotov has been working on designing and optimizing Computer Systems for over 20 years, as an entrepreneur, engineer and researcher.
 
In 2020, Dr Lokhmotov founded KRAI to focus on creating ultra-efficient and cost-effective Computer Systems for emerging Generative AI applications. KRAI is a purveyor of premium benchmarking and optimization solutions for AI Systems. The KRAI team leverages decades of expertise in performance optimization and workflow automation for AI Systems to optimize the Total Cost of Ownership (TCO) and to mitigate risks of deploying AI solutions in enterprise.
 
As a founding member of the industry-leading MLPerf benchmarking competition sometimes called "The Olympics of AI", the KRAI team has been acting as an "Olympic coach" to industry leaders such as Qualcomm, Google, Hewlett Packard Enterprise, Dell Technologies, Nebius and other partners, having contributed to some of the fastest and most energy efficient results in the 7 year history of MLPerf. 
 
As an engineering lead and manager at Arm in 2010-2015, Dr Lokhmotov contributed to the world’s first GPU Compute technologies for mobile and embedded devices, including production and research compilers for the Arm Mali GPU series (over 10 billion units shipped to date).
 
Dr Lokhmotov received a PhD in Computer Science from the University of Cambridge in 2008, and an MSc in Applied Mathematics and Physics (summa cum laude) from the Moscow Institute for Physics and Technology in 2004. Dr Lokhmotov serves as External Member to the Faculty Board and Degree Committee at the Department of Computer Science and Technology at the University of Cambridge.

Wednesday 1st October

Time: 1.00pm Location: HUX 308
Title: Breaking into Quantitative Finance at QRT
Speakers: Orry Gooberman (Quant Technologist) & Anu Krishnamurthi (Talent Acquisition)

Discover what it’s like to start a career in Quantitative and Systematic Trading with Qube Research & Technologies (QRT). This session will provide valuable insights into the industry, life as a graduate, and the opportunities available to you. Hear directly from one of our Quantitative Technology Analysts as he shares his journey from university to QRT, reflecting on his internship, graduate experiences, and how he translated his academic background into a successful career. You’ll also meet one of our campus recruiters, who will walk you through our internship programmes, the unique culture and opportunities at QRT, and how you can apply!

Wednesday 1st October

Time: 2pm Location: HUX 308
Title:  Want to found a start-up? Work at one first!
Speaker:  Lawrence Jones (Engineer at incident.io and DoC alumnus)

The lecture will explore why aspiring founders should first consider working at a high-growth start-up before building their own company. Drawing on first-hand experience at GoCardless as it scaled from 40 to 700 employees, our engineer, Lawrence, will share lessons from navigating the highs and lows of start-up life, taking on broad agency and responsibility, adapting to processes that evolve with growth, and learning how to hire and build strong teams. The talk will also highlight the technical realities behind this journey - from major system migrations and product integrations to designing scalable engineering processes - offering a practical view of how these challenges shape engineering careers.

Tuesday 7th October

Time: 1.00pm Location: HUX 308
Title:  Uncountable -- lessons from starting and growing a company with a machine learning product
Speaker:   Jason Hirshman, Co-Founder

Learn how Uncountable started by using bayesian optimization to recommend experiments to scientists and pivoted to build a growing enterprise software company.

Thursday 9th October

Time: 1pm Location: HUX 308
Title:  How We’re Building an Intelligence Moat at Tuza
Speaker:  Olivia Stannah

In this talk, I’ll share how we’re embedding intelligence into our product to create defensibility in a commoditised market. From proprietary data to learning feedback loops, I’ll explore how we’re building a system of intelligence that gets smarter over time.

Speaker Bio:  I’m Liv, founder and CTO of Tuza—a fintech startup building an intelligent digital store for banks selling card payment solutions. Think Shopify, but for payments.

Tuesday 14th October

Time: 1.00pm Location: HUX 308
Title: JIT Happens: Unleashing the JVM for Low-Latency Trading 
Speaker: Daniel Freeman, Senior Software Engineer at Arctic Lake 

Conventional wisdom suggests that for low-latency code, Java is the wrong tool for the job. So why do many financial firms build their trading systems on the JVM? This talk dives head-first into the myths and realities of high-performance Java. We’ll explore how to tame the Garbage Collector for pause-free execution, and how the Just-In-Time compiler is able to generate machine code that rivals C++. You’ll learn about real-world techniques like zero-allocation programming and off-heap memory, proving that you can have both elite performance and enterprise-grade productivity. 

Thursday 16th October

Time: 1.00pm Location: HUX 308
Title: Innovative Technology in Financial Services

Presenters: Karen Rossi (Lloyds Banking Group CIO), Roxanne Spencer (Lead Data Engineer) & Dafydd Watkins (DevOps Engineer)

Lloyds Banking Group is modernising at-pace to deliver excellent products for our customers. Our tech-led strategy is allowing us to solve new and old problems in dynamic & innovative ways. We will bring that to life with a couple of examples of how we are leveraging Open Source technology & Google Cloud-based automated alerting.

Tuesday 21st October

Time: 1.00pm Location: HUX 308
Title:  From NginX to Rust: The Story of Cloudflare's CDN
Speaker: Maurizio Abba, Senior Engineering Manager

What began as the 'Project Honey Pot' anti-spam system evolved into Cloudflare, a service designed to secure and accelerate any website online. To deliver on that promise, our first CDN was built on NginX with a core focus on low latency. While we added services like Workers over the years, by 2025 that foundational stack was holding us back. To become the fastest CDN on the planet, we rebuilt everything in-house using Rust. This is how we did it."

Thursday 23rd October

Time: 1.00pm Location: HUX 308
Title:  Data Science, AI/ML in Banking
Speaker:  Sabina Kuzina

Lead for Data Science platform and AI/ML Training curriculum for Deutsche Bank, Sabina Kuzina, will present an overview of data science applications and core skills in Investment Banking and Beyond. Frequent speaker at the industry events and conferences on topics of AI and Data Science, Sabina is also a diversity advocate leading Women in AI community globally. 

Tuesday 28th October

Time: 1.00pm Location: HUX 308
Title: The Decade of Agents: Building Agentic AI Applications in Production
Speaker: Zinzan Gurney
 
Agentic AI is moving fast. In this talk, we will map the state of agents in 2025 and what it takes to deploy reliable, trusted systems in production. Examining agents from an engineer’s perspective, we will discuss hard truths such as demo-to-product gaps, hallucinations, evaluation and how AIM Reply is addressing these challenges along with case studies.

Thursday 30th October

Time: 1.00pm Location: HUX 308
Title: Platform Engineering Evolution Through AI @ Cisco
Speaker:  Hasith Kalpage

Find out how Agentic AI was applied at Cisco’s incubator for up-levelling platform engineering with AI, with the effort evolving to an open-source project, caipe.io (Community AI Platform Engineering).

Thursday 6th November

Time: 1.00pm Location: HUX 308
Title: Being human on a dead internet
Speaker: Anastasios Andronidis, Senior Staff Engineer at Tools for Humanity

The internet is dying — not in traffic, but in trust. AI agents, bots, and deepfakes now outnumber humans online, eroding the foundation of what it means to be “real.” In this talk, we’ll explore World ID, a new global identity protocol that proves humanness without revealing who you are. Built around the Orb, a secure hardware device that verifies you’re a unique human, World ID aims to rebuild the web’s social and economic layer on truth instead of probability. We’ll dive into the cryptography, hardware, and AI systems behind it, and discuss what “being human on a dead internet” might mean for the next decade of computing.

Tools for Humanity (TFH) is an information technology company based in Berlin, Germany. Founded by Sam Altman, it is the developer of Worldcoin.

https://www.toolsforhumanity.com/

'