# Deliveroo Deliveroo Tech blog This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## Unlocking AI innovation through automated governance DevFeed: [Unlocking AI innovation through automated governance]() Original publisher: [Read original article]() Author: Alex Evenson Published: 2026-02-27T00:00:00Z Content type: article Language: en Sources: [Deliveroo]() Topics: [Artificial Intelligence](), [AI Development](), [AI Agent](), [MCP](), [Orchestration](), [SDKs](), [agent observability](), [observability](), [tracing]() Tags: [agentic-ai](), [ai](), [ai-development](), [automated](), [governance](), [mcp](), [observability](), [orchestration](), [sdk](), [tracing]() ### AI overview Deliveroo describes building centralized AI infrastructure to combine innovation with governance and compliance. The article introduces its AI Agent Platform and AI Hub, including MCP integrations, workflow orchestration, model-provider abstraction, evaluation pipelines, agent tracing, storage, visualization, and configurable trace retention policies. ### Source excerpt Over the past two years, we've witnessed a Cambrian explosion of AI development, with Generative and Agentic AI capturing stakeholders' attention. AI/ML Engineers are delivering real business value by intermixing LLMs and agentic systems with traditional machine learning systems. It's been an era of rapid prototyping and quick integrations. Teams have adopted different approaches - each unlocking new potential, but also introducing complexity. As these prototypes evolved into production systems, duplicated effort and inefficiencies began to show. Teams built similar observability and reporting solutions in parallel, governance efforts between engineering and legal teams rapidly increased in complexity, and compliance processes developed as bespoke solutions on a per-team basis. We realised that this complexity was limiting our efforts to scale AI projects across the company. To move fast and stay compliant, we needed consistent governance, shared tooling, and clear accountability across teams. We therefore built the components of our AI ecosystem around a central infrastructure - designed for both innovation and oversight. Built upon the principles of automation and transparency, this approach enables our engineers to build safer systems, follow streamlined governance processes, and ultimately ship features faster. Introducing our AI Agent Platform and AI Hub Our in-house AI Agent Platform deserves a blog post of its own, so we will keep it short for now. It is our foundation for developing, deploying, and operating agentic AI systems at Deliveroo. It includes MCP server integrations, handles workflow orchestration, abstracts over different model providers and includes an evaluation SDK for both offline and online eval pipelines. Our AI Agent Platform also gives our engineers full AI agent tracing out of the box, with data storage and user interface components for visualisation. Teams can configure their own data retention policies for these traces to comply with na ## Building AI Agent Applications: Lessons on Model Size, Context, MCP, and RAG DevFeed: [Building AI Agent Applications: Lessons on Model Size, Context, MCP, and RAG]() Original publisher: [Read original article]() Author: Mai Lubega Published: 2026-02-16T00:00:00Z Content type: tutorial Language: en Sources: [Deliveroo]() Topics: [AI Agent](), [Large Language Model](), [Model Context Protocol](), [Retrieval-Augmented Generation](), [Ollama](), [Go Language]() Tags: [ai](), [ai-agent](), [go](), [llms](), [mcp](), [ollama](), [rag](), [software]() ### AI overview This tutorial shares lessons from an Ardan Labs workshop on building AI agent applications. It covers using smaller local models with Ollama, managing conversation context, generating vector embeddings, building RAG applications, and creating an MCP client and server. ### Source excerpt LLMs, MCPs, RAG. There are lots of acronyms in the AI space, but what do they all mean? Dear reader, despite being a software engineer who works in the machine learning space, I confess there was a time I wasn't really sure. Fortunately, with the financial support of Deliveroo's Women-in-Tech Employee Resource Group, I took the Ardan Labs Building AI-Powered Applications in Go workshop that helped me understand what's really going on behind the chat interface and where software engineering meets LLM-based applications. We went through a series of modules to incrementally build a RAG (Retrieval-Augmented Generation) AI Agent application. I started with generating vector embeddings from text and eventually built an MCP Client & Server able to process image and text input and respond to simple queries, after which I felt I understood the new AI landscape much, much better. For my dearest gentle reader, I'm happy to share these three takeaways from the course: Size Matters Full-size language models are very large and can require several hundred gigabytes of memory to run, certainly much more than is available on a personal laptop. To work around this constraint, we used Ollama -- an open-source project that provides smaller, more size-efficient LLM models - that can be run locally. In the real world, some applications get around this problem by sending data directly to OpenAI/Anthropic's APIs. They don't have to host any models locally, just send and receive data (subject to usage tier and API limits, of course). But applications that need a model to run locally for proprietary reasons will have to manage the infrastructure themselves -- which usually involves a very spicy cloud bill or a server in the corner that doubles as a space heater. Don't Lose the Plot (Manage your Context) Think of Context as the LLM's short-term memory. When we chat with an AI, we don't just send the current question; we send the entire conversation history, system instructions, and any retrieve ## Engineering at Deliveroo India: How We Build, Collaborate, and Grow DevFeed: [Engineering at Deliveroo India: How We Build, Collaborate, and Grow]() Original publisher: [Read original article]() Author: Kartik Visvanathan Published: 2025-11-06T00:00:00Z Content type: article Language: en Sources: [Deliveroo]() Topics: [Development](), [idc](), [Slack](), [Figma](), [systems]() Tags: [asynchronous](), [collaboration](), [cross-functional-teams](), [development](), [engineering](), [figma](), [india](), [slack]() ### AI overview Kartik Visvanathan describes how Deliveroo's India engineering hub operates as a global, cross-functional development center. The article covers team ownership, collaboration across time zones, co-located roles, asynchronous communication, and tools including Slack, Confluence, and Figma. ### Source excerpt Hey everyone, I'm Kartik Visvanathan, an Engineering Manager at Deliveroo, based in India. I joined the team in 2023, and it's been an exciting ride ever since. I've built some truly impactful products here, including an advertising platform that we scaled into a profitable business, and a corporate orders product that touched every part of our systems--from restaurants to delivery to customer service. But while building cool stuff is awesome, what I really want to share today is the secret sauce behind it: how we work as a team here in India. A Truly Global Hub Our engineering hub in India is a core driver of Deliveroo's global mission. Over the past few years, our teams have transitioned and taken complete ownership of key product areas. Whether we're improving the restaurant onboarding flow or fine-tuning our delivery times, the work we do here is absolutely core to Deliveroo's success. When our India Development Center (IDC) started a few years back, we had engineers reporting directly to managers in the UK. I was one of the first engineering managers hired in India in 2023, and I was given the opportunity to build a team that was entirely based here. We've since grown to include other roles like product managers, analytics engineers, and data scientists right here in India. This lets us have "co-located teams"--people who can sit in the same room, grab a whiteboard, and even go out for a team social that isn't virtual. When candidates ask about my favorite parts of the job, I always mention three things: cross-team collaboration, work flexibility and life harmony, and a culture of ownership. So let's dive into those. How We Collaborate Across Teams Working with teammates in London and other time zones means a lot of asynchronous communication. We rely heavily on tools like Slack, Confluence, and Figma, which are essentially our digital whiteboards. Regular virtual stand-ups and a focus on clear ownership help us build trust and keep things moving. In the advertis ## Cost Optimisation in ECS: Integrating Spot Instances at Scale DevFeed: [Cost Optimisation in ECS: Integrating Spot Instances at Scale]() Original publisher: [Read original article]() Author: Aakash Singhal Published: 2025-09-12T00:00:00Z Content type: article Language: en Sources: [Deliveroo]() Topics: [Amazon Elastic Container Service](), [Amazon EC2](), [Amazon Web Services](), [Cloud](), [Automation](), [Resilience](), [Availability](), [container]() Tags: [automation](), [availability](), [aws](), [container](), [cost](), [cost-optimisation](), [ec2](), [ecs](), [infrastructure](), [load-balancer](), [resilience](), [scale](), [stateless](), [storage]() ### AI overview Deliveroo describes how it integrated EC2 Spot Instances into Amazon ECS to reduce compute costs while maintaining service stability. The approach routes only eligible workloads to Spot capacity and uses criteria such as fast shutdown, task redundancy, statelessness, and load balancer deregistration timing. ### Source excerpt At Deliveroo, we're always refining how we scale - especially when it comes to managing compute costs in the cloud. After optimising our Amazon ECS workloads with Reserved Instances and Savings Plans, we saw an opportunity to push further using EC2 Spot Instances, which offer up to 90% savings compared to On-Demand prices. But Spot comes with challenges: Their availability can fluctuate, and they can be terminated with just a two-minute warning. To unlock these savings without compromising service stability, we had to engineer a robust solution across infrastructure, workload qualification, and automation. The Challenge: Balancing Cost and Reliability Our ECS infrastructure initially relied entirely on On-Demand EC2 instances, provisioned through Auto Scaling Groups (ASGs) connected to ECS Capacity Providers. While reliable, this approach didn't take advantage of AWS's surplus compute capacity. We aimed to layer Spot Instances into our clusters, but selectively. Our goal was clear: route only eligible workloads to Spot capacity while ensuring no service degradation during unexpected terminations. Spot Instances: Power and Pitfalls Spot Instances provide dramatic cost reductions but introduce several operational caveats: Ephemeral by nature: AWS can terminate them at any time with a two-minute warning. Capacity variability: Availability depends on AWS's excess capacity in each AZ and can shift unpredictably. Scaling limitations: Auto Scaling may fail if the desired instance types are not currently available. To avoid introducing fragility into our stack, we established technical eligibility criteria that workloads must meet before being scheduled on Spot. Defining Spot Eligibility We formalised the following criteria to assess whether a workload could safely tolerate Spot interruptions: Fast Shutdown Support Constraint: stopTimeout must be < 120 seconds in the container definition. Reason: Ensures ECS has time to gracefully shut down the task before AWS's 2-minute te ## How Deliveroo Built Its Machine Learning Platform DevFeed: [How Deliveroo Built Its Machine Learning Platform]() Original publisher: [Read original article]() Author: Saikrishna Desaraju Published: 2025-07-02T00:00:00Z Content type: article Language: en Sources: [Deliveroo]() Topics: [Machine Learning & Artificial Intelligence](), [Deployment](), [model-deployment](), [Kubernetes](), [Amazon Web Services](), [distributed-training](), [Inference](), [Python](), [PyTorch](), [Tensorflow]() Tags: [aws](), [deployment](), [distributed-training](), [inference](), [kubernetes](), [machine-learning](), [machine-learning-platform](), [python](), [pytorch](), [rest](), [tensorflow]() ### AI overview Deliveroo describes its machine learning platform for standardizing model-building and deployment workflows. The platform combines open-source tools, in-house components, Kubernetes on AWS EKS, TensorFlow, PyTorch, Metaflow, and Argo; it also supports GPU and distributed training and includes Inferoo for real-time model predictions. ### Source excerpt Enter Deliveroo's ML Platform For the past three years, we have been building Deliveroo's Machine Learning Platform, or the ML Platform as we like to call it. The ML Platform boosts our model-building and deployment capabilities by standardising ML workflows, streamlining the end-to-end development process and simplifying model deployment. Besides saving software engineering effort through centralising tooling, the ML Platform also reduces the time that our ML engineers spend on infrastructure tasks. As a result, our ML engineers can now iterate their ML models 2-3x faster than before. What makes our ML Platform tick? At the core of our ML Platform lies a carefully curated tech stack - an integrated suite of infrastructure tools, services, and libraries. It blends robust open source technologies with purpose-built, in-house components. Key open source tools include Kubernetes, Argo, and Metaflow, all seamlessly connected with leading ML frameworks like TensorFlow and PyTorch. We choose mature, community-driven solutions and actively contribute back where we can. This entire ecosystem is powered by AWS, running on EKS, and anchored by our data warehouse. In the sections that follow, we'll take a closer look at the key components that drive our ML Platform. Metaflow To give ML engineers seamless access to scalable compute on Kubernetes, we use Metaflow, a powerful Python library that helps break down complex model-building workflows into smaller, manageable jobs. These jobs are orchestrated by Argo, one of the backbone tools in our infrastructure. One of Metaflow's biggest strengths is its flexibility. It allows engineers to move effortlessly between local development, staging on Kubernetes, and production deployment--helping teams iterate quickly as they experiment and scale. Recently, we added support for GPUs and distributed training, enabling faster training times and the ability to work with much larger datasets. Inferoo For models that require real-time predictio ## Antonio Troina: My Journey So Far as a Senior Staff Software Engineer DevFeed: [Antonio Troina: My Journey So Far as a Senior Staff Software Engineer]() Original publisher: [Read original article]() Author: Antonio Troina Published: 2025-03-24T00:00:00Z Content type: opinion Language: en Sources: [Deliveroo]() Topics: [Software Engineering](), [engineering-culture](), [Development]() Tags: [advice](), [career](), [individual-contributor](), [software-engineer](), [software-engineering]() ### AI overview Antonio Troina describes his six-and-a-half-year career at Deliveroo, from consultancy work to becoming a Senior Staff Software Engineer. He discusses guiding teams, reframing technical and product problems, and progressing along the individual-contributor path. ### Source excerpt Could you introduce yourself and your role at Deliveroo? I'm Antonio, a Senior Staff Engineer at Deliveroo. I have been with Deliveroo for six and a half years, overseeing the entire New Verticals domain as well as the Catalogues space, as an Individual Contributor (IC). Can you tell us a little bit about your career journey into engineering and how you found yourself at Deliveroo? I have been passionate about computers since childhood, which led me to pursue studies in software engineering. I began my career working for consultancy firms in Italy, my home country, before moving to London, where I continued in consultancy roles across both the private and public sectors. Eventually, I joined Deliveroo, which, interestingly, is my first purely product-based company. I started as a Software Engineer six and a half years ago, and as my skills, experience, and ability to influence grew, I naturally transitioned into my current role, which I truly enjoy. Reflecting on the role of a Senior Staff Software Engineer here, what do you think is the most rewarding part of it? One of the most interesting aspects of my role is helping teams see challenges from different perspectives, drawing from my experience over the years. I've encountered situations where problems were approached in one way but could have been tackled more effectively through alternative methods. I enjoy guiding teams through that journey, particularly when it comes to challenging product requirements and rethinking how problems are framed. While many initiatives originate from product, some also stem directly from technology. Often, our technical interactions with other teams and systems provide insights that may not be as clear to product managers. In these cases, I find that a bottom-up approach can be highly effective. What would you say about career progression at Deliveroo? I would say career progression is quite varied. It can involve moving upward within the same path or making a lateral shift to expl ## Women in Tech presents: Roo Tech Showcase DevFeed: [Women in Tech presents: Roo Tech Showcase]() Original publisher: [Read original article]() Author: Olga Afanasieva Published: 2025-03-13T00:00:00Z Content type: opinion Language: en Sources: [Deliveroo]() Topics: [Software](), [App](), [coding]() Tags: [building](), [community](), [diversity](), [event](), [product](), [projects](), [software](), [talks](), [tech](), [women-in-tech]() ### AI overview Deliveroo's Women in Tech Employee Resource Group presents eight female employees and their five-minute lightning talks about projects, career transitions, and ideas shared at the group's end-of-year event. ### Source excerpt The Women in Tech Employee Resource Group (ERG) at Deliveroo was officially founded in 2023, and we are proud to represent an amazing community of female employees who identify as working in Tech at Deliveroo. This includes, but is not limited to, Software Engineers, Data Scientists, Designers, and Product Managers and we also welcome those who want to move into a Tech role. We host many events and initiatives throughout the year that align with our mission to connect and empower women in technology to excel in their careers at Deliveroo and beyond. For our main end-of-year event, we chose to showcase some of the wonderful projects our female employees achieved last year, in the form of 5-minute lightning talks. Our VP of Engineering, Vaughn Washington, opened the event with a message about the importance of our ERG and events like these, which provide underrepresented groups a forum to showcase their achievements. In continuation of this spirit, we are very excited to be publishing this post on our tech blog for more people to discover our eight remarkable female speakers and their innovative ideas. Enjoy! Olga Afanasieva, Product Manager My name is Olga, I've been at Deliveroo about 3.5 years and mostly looked after Consumer Search and Recommendations. During the event I shared with the team my journey of building a Proof of Concept (POC) of an app that could help couples evaluate their relationship, facilitate feedback exchange and help them nurture the relationship. I've built that POC with my husband and a certified couple therapist while I was on maternity leave and managed to test it with 15 couples. Rabia Lakhani, Product Manager I'm Rabia and I am a product manager in the Consumer New Verticals team looking after the browsing, upsell and basket experience for our customers. I joined Deliveroo as a Technical Programme Manager and around 12 months ago, started my official journey into the Product world. I wanted to share my personal transition story on how I ## Gina Wiley: My Journey So Far as a Senior Engineering Manager DevFeed: [Gina Wiley: My Journey So Far as a Senior Engineering Manager]() Original publisher: [Read original article]() Author: Gina Wiley Published: 2025-02-20T00:00:00Z Content type: article Language: en Sources: [Deliveroo]() Topics: [engineering-culture](), [Human-Computer Interaction and Visualization](), [Shared Responsibility Model]() Tags: [career](), [culture](), [engineering](), [engineering-culture](), [experience](), [growth](), [human-computer-interaction](), [journey](), [saas](), [team]() ### AI overview Gina Wiley describes her path from studying information systems and human-computer interaction to engineering management at Deliveroo. She discusses consulting, DoorDash, Deliveroo's New Verticals team, and her views on supporting engineers through ownership, resources, and clear career expectations. ### Source excerpt Could you introduce yourself and your role at Deliveroo? I'm Gina, I'm a Senior Engineering Manager at Deliveroo. I lead our Consumer Experience group in our New Verticals engineering team, and I've been at Deliveroo for five years. Could you walk us through your path to engineering management and what brought you to Deliveroo? I studied information systems at university and developed an interest in human-computer interaction. This field emphasises understanding users and translating business processes into technical requirements. I moved to San Francisco for a tech consulting role, where I worked on a large-scale, multi-country project replacing legacy systems with connected Software as a Service (SaaS) products. This cross-functional experience shaped my understanding of end-to-end business process transformation through technology. In 2017, I joined DoorDash's Business Applications team, working on partner onboarding and support processes, which introduced me to logistics and delivery tech during a pivotal time for the industry. In 2020, I moved to London to work at Deliveroo, initially supporting restaurant partner onboarding in the Restaurants engineering group. I now work in the New Verticals engineering group, focusing on launching grocery and retail propositions. We've recently introduced a lot of new features to support this growing space; so continuing to innovate and deliver at pace to keep up with the changing market has been an exciting challenge! Reflecting on the role of engineering manager here, what do you think is the most rewarding part of this role? I enjoy seeing engineers on the team take on new challenges and push themselves out of their comfort zone to reach their full potential, so I'll make sure that they get access to the right resources and opportunities so that they can continue to learn and grow. It's important to trust your people to reach their growth goals with the right opportunities. I believe (and have seen throughout my career) t ## Deliveroo experimentation principles DevFeed: [Deliveroo experimentation principles]() Original publisher: [Read original article]() Author: Ella Johnson-Watts Published: 2025-02-06T00:00:00Z Content type: article Language: en Sources: [Deliveroo]() Topics: [A/B Testing](), [experiments](), [Testing](), [Tooling]() Tags: [a-b-testing](), [algorithm](), [development](), [testing](), [tooling](), [ui]() ### AI overview Deliveroo describes how its experimentation programme evolved from ad hoc practices into a mature platform and culture. The article introduces principles such as defining a clear hypothesis and success criteria before running an experiment. ### Source excerpt Experimentation at Deliveroo has significantly boosted our innovation as a company. It has allowed us to explore new ideas, protected us from unforeseen challenges by preventing wasted time and resources on ineffective initiatives, and promoted intellectual honesty by validating our assumptions. Over time, the breadth of our experimentation programme, our tooling and our culture have undergone significant development. Each year, we run hundreds of experiments, testing a wide array of features. These experiments range from simple UI modifications to complex algorithm adjustments. We employ a variety of methods, from standard A/B testing to advanced techniques like interleaving, multi-armed bandits and switchbacks. We've moved from teams using ad-hoc methods for deploying and analysing experiments to a mature experimentation platform, which standardises and automates much of the experiment process. But successful experimentation requires more than just technical expertise and tooling. It requires an experimentation "culture" - a shared set of principles and values that embeds experiments in the broader organisational context, and ensures that experimentation is done in a way that is maximally impactful. As we have matured technically and increasingly democratised our technical capabilities, we have also evolved a set of such principles, which codifies that cultural side. These enable us to uphold high-quality experimentation whilst achieving our current goal of broadening the pool of Deliveroo employees capable of running experiments. Principle 1: Every experiment starts with a clear hypothesis and success criteria. Experiments work best when they are specific. We formulate our hypothesis based on past experiments, anecdotes, user research, and competitive analysis - leveraging existing learnings helps us design better experiments. We avoid experimenting aimlessly just to 'see what happens', risking learning nothing. Experiments are a highly valuable resource and unst