Informations
Jobs
Team Inc(AI)00AIData scienceWorking methodsAI Agents in Retail: How GAIA Automates Decision-Making Processes at OTTO
With the GAIA Cockpit, OTTO is developing a control tool for the AI agent GAIA. Using the GEMA meeting as an example, this article shows how AI agents help reduce the effort required for preparation and follow-up, consolidate information from various sources, and facilitate faster, well-informed decision-making. The technical foundation consists of Snowflake Cortex AI and a custom-developed Streamlit app.
Falk024DevelopmentOperationsCloudRewriting OTTO in Rust: oxidizing our codebase, one service at a time
Why are we rewriting OTTO in Rust? ✍️💻 Find out why we’ve chosen Rust and how we’ve migrated Lambdas and microservices.👉 Read more here.
Nina07AIDevelopmentFrom Prompt to Product: Spec-Driven Development with Agent Workflows
Spec-Driven Development at OTTO: How to build a frontend without vibe coding and without manually written code 👉 Read more now!
Andreas00DevelopmentCloudSecuritySecure Service Tokens with Keycloak Token Exchange – Automated with Terraform
🖥️ Learn how to set up suitable service tokens for microservices with Keycloak Token Exchange – automated with Terraform 👉 Read now!
Jan09AIDevelopmentAI Software Development –How AI Agents Are Transforming Our Engineering
How we at OTTO are re-engineering AWS microservices using AI agents and Git-based architecture 👉 Read now!
Birgit08AIWorking methodsWhy AI Hallucinates – and What Technical Writing Can Do About It
Why AI systems hallucinate and how technical writing can help achieve more accurate results through structure, terminology and metadata. ➡️ Read now!- Nina14Data scienceDevelopmentCloud
Terraform Tips: Securing Infrastructure-as-Code with Terratest – Part 3
Learn how to test Terraform modules with Terratest, detect errors early on and reliably secure IaC in CI/CD pipelines. 🖥️✔️ 
Aemilia und Svenja12Data scienceWorking methodsData-Driven Team Development – How OTTO Leads Tech Teams to Success
What makes tech teams truly effective? How we use data-driven team development to measure effectiveness. ➡️ Explore the blogpost
Team Jarvis04DevelopmentMachine learningLearning to Rank is Rocket Science: How Clojure accelerates Our Machine Learning with Deep Neural Networks
Learn how OTTO uses 🖥️ Clojure to build robust machine learning pipelines for learning to rank 🚀 and improves the search experience.- Team Jarvis01Development
Learning to Rank – To the Moon and Back(-propagation): Deep Neural Networks That Learn to Rank What You Love
Learn how OTTO optimizes search relevance using Learning to Rank – and the role Deep Learning plays for our Data Science team. 
Marco02DevelopmentOpen Source at OTTO: Transparency, Community & Innovation
Find out how OTTO lives Open Source – for more transparency, developer satisfaction and digital sovereignty. Read more about our strategy now!
Aashrika01DevelopmentEvents and cultureWorking methodsAWS & GCP – Two Summits, Countless Insights
Our personal insights & highlights ☆ AWS Summit Hamburg & GCP Summit Düsseldorf- Nina02DevelopmentWorking methodsMachine learning
What is an LLM? | Large Language Models explained
Learn how large language models (LLM) like ChatGPT work. An overview of the structure, training & potential of modern language models. - Marvin01Development
Visualizing Software Architecture with C4 Model
Problems with architecture diagram creation using traditional methods – e.g. UML? Solution: C4 model ► Tips for implementation & use - Andreas00Development
How UV Solves The Hard Problem of Python’s Package Management
Package management in Python ✓ Master challenges with UV package manager ✓ Read more about organization & benefits now! 
Alexander01AIData scienceMachine learningAI assistant: From Vector Databases to Multimodal Agents
Optimization of our AI assistant ✓ Use of agent-based modeling ✓ Read more about reasons for optimization & the path now!- Nina02DevelopmentWorking methods
Terraform Tips: Scalable and Reliable Infrastructure-as-Code – Part 2
Optimize directory structure & resource organization ✓ Handling of input & output values ✓ Read more about Terraform Best Practices now! - Nina01DevelopmentWorking methods
Terraform Tips: Scalable and Reliable Infrastructure-as-Code – Part 1
Terraform Best Practices ✓ Define Infrastructure-as-Code ✓ Terraform Lifecycle & Terraform Modules ✓ Read more now! 
COST01DevelopmentCloudHow we used a simple trick to save USD 500,000 in data transfer costs
Increased efficiency of cloud usage ✓ Cost reduction ✓ Migration to dual-stack IPv6 ✓ Find out more now!
Dominik01DevelopmentWorking methodsDeveloper Hacks – Modern Command Line Tools and Advanced Git Commands
In this article, you learn more about a modern development setup, state-of-the-art alternatives to classic shell programs, and advanced git commands.
Team PIT01ArchitectureDevelopmentOperationsAbout the development of genAI assistants AskARev and Searchbuddy
Find out (in a very special way) how the PIT team at Otto Group data.works developed the generative AI-based products AskARev and Searchbuddy.- Nina00DevelopmentOperationsWorking methods
Meet Nina – From Healthcare Management to Software Development
Find out how Nina mastered her career change into software development with the help of her team and her great enthusiasm for IT! 
Fabian00DevelopmentOperationsWorking methodsConfluent Helps Power a Diverse and Scalable Online Shopping Experience for OTTO
Find out how OTTO is taking its 10-year-old data infrastructure into a new era with the event streaming platform powered by Confluent and making itself future-proof.- Jim00Data scienceDevelopmentMachine learning
Machine Learning Ops: What It's Like in Practice
Get an insight into Jim's work as an MLOps-focused engineer! In this article, he explains how a machine learning product is developed and how you can get started in this discipline yourself. - Mirco00Data scienceDevelopment
Spanner's Distributed Nature
At OTTO, Google Cloud Spanner is an integral part of several services. Mirco reveals the complexity of optimizing joins in Google Cloud Spanner and gives some helpful insider tips. - Falk00DevelopmentEvents and culture
Recap of the EuroRust 2023 conference in Brussels
Kay and Falk were at the EuroRust conference in Brussels in mid-October. There are more and more enthusiasts of the Rust programming language at OTTO. The two of them tell you what went on at the conference and why Rust is a useful tool. 
Daniel00DevelopmentSecuritySecurely managing sensitive data with HashiCorp Vault
Almost every application is using sensitive data, such as passwords or other login credentials, to log in to databases, third-party systems and the like. Such data should not end up in plain text in the version control system, but rather be kept under lock and key. For the secure storage of data, the use of Hashicorp Vault has gained acceptance at OTTO.
Christian00DevelopmentWorking methodsDagster: Data Engineering Best Practice during Slack Time
In the inc(AI) team, we are always trying out different tools and methods. Slack time has proven to be a good way of doing this, in which we learn alone or together. The motivation for such Slack time naturally comes from the challenges that we have to solve during our work. Our strategy is to always look to the open source community first to find suitable solutions for us. Some time ago, we also investigated a new tool called Dagster in our Slack time.- Stefanie00ArchitectureDevelopmentWorking methods
How to setup a Kanban board in Jira using sub columns
In this blog post I want to share the evolution of our Kanban board in Jira and how we worked around the missing feature of sub columns. - Christian00DevelopmentMachine learning
ProductGPT – automating the creation of product texts with ChatGPT
Although they've been around for a while, Large Language Models (LLMs) have taken the world by storm since ChatGPT was launched. ChatGPT currently dominates the AI landscape like no other model. Could it also be used for the creation of good product texts in the OTTO Webshop? This is precisely the question we wanted to tackle in the inc(AI) team in a 2-week sprint. The result is ProductGPT – a tool that can do exactly that! - Tobias00Data scienceMachine learning
User centricity beginning with the first keystroke
The majority of user journeys in search starts with a click in a search box. How can you guide your users after this initial click? At otto.de we strive to show our users the best suggestions for their need, right after the first click - Birgit00ArchitectureDevelopmentWorking methods
Why using API Guidelines is a good idea
Working according to a standard... really? Come on – software developers know how to build APIs. Besides, they want to stay flexible and have neither the desire nor the time to plough their way through some endless set of rules. And even if you've read them, that doesn't mean you agree with them. Fair enough. Nevertheless, there are a few challenges to overcome in software development, especially when working in autonomous teams and when you're dealing with business growth. This post offers an insight into how we at OTTO came up with our own API Guidelines and implemented them successfully. 
Alexander00DevelopmentCloudSecurityEnterprise-scaled Self-Healing StackSets
At OTTO, we need to manage several hundred AWS accounts for our product teams while balancing agility and control. That's why OTTO IT's Cloud Competence Center, also known as the Governance at Scale (GAS) team, worked with globaldatanet to develop a solution to run AWS CloudFormation StackSets at Scale, integrated into the OTTO tooling ecosystem with Confluence and Microsoft Teams.- Philipp00DevelopmentCloud
Infrastructure as Code (IaC) – keeping Terraform configuration DRY with Terragrunt
The key is to minimize, or even completely eliminate, code repetition by keeping the Terraform configuration DRY (Don’t Repeat Yourself). This is why we migrated to Terragrunt. It's a small wrapper around Terraform promising reduced repetition, better remote-state configuration, and simpler work with Terraform modules. It can be a viable solution for better organizing configuration files and clearly stating differences between deployment environments. 
Techblogger00Data scienceDevelopmentWorking methodsYesWeHack and OTTO stage Live Bug Bounty Hunting event
At this year's Nullcon Berlin, the international IT security conference that brings hundreds of top experts together, we staged a live Bug Bounty Hunt together with YesWeHack, a leading provider of bug bounty services. Around 40 crack security researchers pitched in to subject OTTO's infrastructure security to a hardcore test. This was a welcome chance for us to check our Web applications for security loopholes and learn from a personal exchange with leading-edge researchers at the same time.
Christian01Data scienceWorking methodsReturn-rate forecasting with survival analysis
Especially in OTTO's retail business, accurately forecasting return rates and their fluctuations plays a major role in minimising unnecessary costs. In large-goods deliveries (fridges, sofas etc.) every single return results in very significant costs.
Michael00Development'Real' real-time programmatic advertising with TensorFlow
In collaboration with Online Marketing, Team Siggi at OTTO BI developed an in-house programmatic advertising solution in 2018. For this purpose, among other elements a separate SSP (Supply Side Platform) and DSP (Demand Side Platform) solution was developed, called Orbidder.
Team Nautik01DevelopmentA Machine Learning approach to real-time display advertising Click here – it’s not an ad, we promise ;-)
Let’s be honest, display advertising usually makes people think of an aggressive e-tailers literally hunting them down. Many potential online customers therefore resort to all-or-nothing ad blockers to save themselves from visual overload. However, beyond these two extreme user experiences, in reality display advertising is simply a modern equivalent to a classic poster hanging at the bus stop.- Techblogger00Development
Rethinking semantic product search: a new approach to queries with unusual spelling variations
In recent years, semantic product search has been widely adopted in global e-commerce. However, many systems exhibit issues in understanding queries with unusual spelling variations. In this blogpost, we show you how Otto’s Team Turing has developed a new system that alleviates these problems. It is live and has lifted the Click-Through Rates (CTR) in the low-recall segments of otto.de by more than 25% and the Conversion Rate (CvR) in the low-hit segment by more than 90%. 
Andreas00DevelopmentKaggle: Machine Learning competition for the best Multi-Objective Recommender System
Major online retailers such as OTTO offer their customers millions of products to explore and purchase. However, finding the right product from such a vast selection without a little guidance can be exhausting! So we work to guide our customers to those products that best match their interests and motivations using personalised recommendations. For this reason we want to enhance our ability to forecast in real time which products each customer will want to see, add to their cart and order at any given moment of their visit.
