CWDN series: Dev-eXperience - DataStax: Meet developers where they want to be
This is a guest post for the Computer Weekly Developer Network written by Dom Couldwell in his capacity as head of field engineering for EMEA region at data platform company DataStax.
DataStax is the company behind the scalable, highly available, cloud-native NoSQL database built on Apache Cassandra – the company is focused on enabling developers to build real-time apps that connect data at rest and in motion across multi-cloud deployments and it has a vested interest in open source enterprise technologies.
Couldwell writes in full as follows…
Making developers happy used to be about the quality of snacks and how good the office Internet connection was and it was viewed as a nice-to-have.
Today, it’s business critical.
According to Gartner, 58 percent of software engineering leaders think developer experience (DX) is ‘very’ or ‘extremely’ critical to the business leaders at their organisations. Why? Because developers build the software that powers that business.
DX is about helping developers build new applications, or make necessary changes to existing systems. Whether it is adding new functionality or fixing performance issues, DX is about getting out of the way as much as possible. By making it easier for developers to do what they need, feature velocity increases and the business benefits.
In order to improve DX, we have to know how developers work today and where they want to get to. This includes some of the typical technical choices that developers and architects have to make, like which APIs or programming languages to standardise on, but it also has to look forward too.
Where do you want to go today?
With so much discussion around AI and automated agents going on, delivering good DX involves looking at what innovations are going on and how those new launches can be integrated into existing services. Essentially, providing good DX means being in more places where developers want to be and supporting or integrating with those new projects. You may have the best product on the market but if it’s not easy for developers to access your product from their favourite tools, adoption will be hampered.
Couldwell: We’ve gone beyond ‘nice-to-have’ with Dev-eX today/
For developers jumping into the world of AI, it can be especially daunting. A plethora of new terminology and tools is allied with ever more demanding data needs for events, features, predictions and the pipelines that drive. To make DX better here, developers need help to get over the hurdles and integrate with the tools that they are familiar with.
So where can you get started and improve DX?
To meet developers where they are today, we have to make the tools required to build AI-powered experiences as familiar as possible and take away the pain of having to learn a whole new set of skills. Yes, developers want to get up to speed with the latest tech, but they also need to deliver in short timeframes.
A typical developer’s need might be to build a new service with AI, based on Retrieval Augmented Generation or RAG to supplement an existing Large Language Model (LLM) with new data created over time. Implementing this could involve multiple interactions with a vector database, an LLM and a database containing additional metadata. Connecting all these services needs to be as seamless as possible. For example exposing interfaces through APIs and consolidating the data storage into one flexible location where developers can quickly upload test data, vectorise and retrieve it as needed.
This accelerates the initial development stage of application development, as well as making DX easier over time.
From test to production
When taking the app you have developed into production, you face a different set of challenges. Rather than exporting data to that new AI service or Large Language Model, we should bring the AI service or LLM to the data that we already have. This solves two problems. Firstly, it makes it easier for the organisation as a whole to stay in charge of its approach to any new technology. Secondly, it keeps developers working in environments that they are already familiar with and know well.
AI is becoming more and more commoditised every day, be it reducing computing needs, simplifying the process of training models or making every database vector enabled. For developers building AI-powered applications, we need the same level of simplification applied to DX.
