Level Up Your Developer Experience (DevEx) with GenAI

Ai Generated Mushrooms Psychedelics royalty-free stock illustration. https://pixabay.com/

What is Developer Experience (DevEx)?

DevEx encompasses the array of systems, technology, processes, and cultural aspects that impact the efficiency of software development. It examines every facet of a developer’s environment, including firm’s culture, workflows, tools, and overall ecosystem, to determine their collective effect on developer productivity, contentment, and operational outcomes.

GitHub’s DevEx formula takes into account:

• Productivity: how quickly or simply a change can be made to a codebase
• Impact: how frictionless it is to move from idea to production
• Satisfaction: how the environment, workflows, and tools affect developer happiness

C = Collaboration, the multiplier across the entire DevEx.

Optimizing DevEx is about creating a collaborative environment where developers can be their most productive, impactful, and satisfied.

Regrettably, there are presently no universally recognized industry standards for gauging DevEx. Nonetheless, Mondello suggests leveraging the DevOps Research and Assessment (DORA) framework, which assesses an organization’s DevOps performance, as a valuable resource. This framework includes key metrics such as:

1. Deployment frequency (DF): The frequency at which an organization releases new software.
2. Lead time for changes (LT): The duration from when a change is requested or initiated to its deployment.
3. Mean time to recovery (MTTR): The average duration to recover from a failure.
4. Change failure rate (CFR): The percentage of changes resulting in failure.

With DevEx and GenAI, every aspect of a developer’s journey is questioned.

Specific benefits GenAI offers to improve DevEx

GenAI significantly enhances DevEx by streamlining workflows, automating repetitive tasks, and providing actionable insights. It boosts developer productivity, accelerates time-to-market, and fosters innovation by empowering teams to focus on high-value activities. Moreover, GenAI minimizes errors, enhances collaboration, and optimizes resource utilization, leading to more efficient and effective software development processes.

Let’s explore a few use cases and examples of how GenAI impacts DevEx.

1. Coding Assistant

AI coding assistant is a tool that uses GenAI to help you write code more efficiently and accurately. It’s like having a super-smart code reviewer, autocomplete, and integrated chat tool all in one.

There are already a few paid and free players in the market, including: Codeium, Blackbox AI, GPT4ALL, Tabby, GitHub Copilot, Tabnine and my personal favorite, JetBrains AI.

2. Code Generation

OpenDevin, a revolutionary autonomous agent, is engineered to navigate the intricacies of software development with precision. Integrating tools like a shell, code editor, and web browser, it demonstrates the vast capabilities of LLMs in this domain. Stemming from the aspiration to replicate, refine, and push the boundaries of the original Devin model, the OpenDevin project is a testament to innovation. Through collaboration with the open-source community, it endeavors to address real-world challenges encountered by Code LLMs, fostering impactful contributions and laying the groundwork for future progress.

Check out my recent experiments with OpenDevin. I anticipate that organizations will soon embrace and refine the model, enhancing their code generation capabilities to empower their developers. This shift will enable developers to transition from mundane tasks to accelerated development, driving efficiency and innovation.

Chris Shayan on LinkedIn: #opendevin #innovation #collaboration #techcommunity #llm #llmcodegen

3. Generating Test Data

A common challenge in development is creating sufficient test data efficiently to facilitate Test-Driven Development (TDD) and Behavior-Driven Development (BDD), thereby enhancing software quality.

Check out my latest experiment automating Test Data Generation for Vietnamese Banking Customers using Llama & Ollama. Are you facing challenges creating realistic test data tailored to your Vietnamese customer base? Llama & Ollama have you covered! In this brief demonstration, I showcase how you can harness Llama’s data generation features alongside Ollama’s Vietnamese address library to generate a series of sample customer profiles. This dataset proves invaluable for testing diverse functionalities within your application, guaranteeing a smooth and optimized experience for your Vietnamese users.

Construct the JSON schema for your desired test data:

template = {
    "firstName": "",
    "lastName": "",
    "address": {
        "street": "",
        "city": "",
        "country": "",
        "zipCode": ""
    },
    "phoneNumber": "",
    "account": {
        "account_number": "",
        "balance": "",
        "transactions": [
            {
                "dateTimeHHMMSS": "",
                "description": "",
                "amount": ""
            }
        ]
    },
    "creditCardNumber": "",
    "creditCardLimit": "",
    "debitCardNumber": ""
}

Create your prompt:

prompt = (f"generate one realistically believable sample data set of a persons first name, last name, address in the "
          f"Vietnam, phone number, current account number, balance for current account in VND, transaction detail (at least 10 records) "
          f"including date and time, description (deposit, withdraw or interest), realistic VISA or"
          f"Mastercard credit card"
          f"number, limit for credit card using VND, and a VISA or Mastercard debit card number. \nUse the following "
          f"template:"
          f"{json.dumps(template)}.")

And there you have it: a sample of realistic banking customer data in Vietnam. It’s worth noting that the model understands that withdrawals imply negative amounts without the need for explicit instruction.

{
 "firstName": "Anh",
 "lastName": "Truong",
 "address": {
 "street": "Nguyen Du Street",
 "city": "Hanoi",
 "country": "Vietnam",
 "zipCode": "10000"
 },
 "phoneNumber": "+84 931 725 842",
 "account": {
 "account_number": "1234567890",
 "balance": "1000000",
 "transactions": [
 {
 "dateTimeHHMMSS": "2023-02-20 10:00:00",
 "description": "Deposit",
 "amount": "500000"
 },
 {
 "dateTimeHHMMSS": "2023-02-15 12:00:00",
 "description": "Withdraw",
 "amount": "-20000"
 },
 {
 "dateTimeHHMMSS": "2023-02-18 14:00:00",
 "description": "Interest",
 "amount": "500000"
 },
 {
 "dateTimeHHMMSS": "2023-02-21 09:00:00",
 "description": "Deposit",
 "amount": "250000"
 },
 {
 "dateTimeHHMMSS": "2023-02-22 16:00:00",
 "description": "Withdraw",
 "amount": "-30000"
 }
 ]
 },
 "creditCardNumber": "5104111123456789",
 "creditCardLimit": "100000",
 "debitCardNumber": "62543210112345678"
}

Picture this: all your microservices are exposed on your API Gateway (you can refer to this post for more details on this desired architecture). If you’ve been leveraging APIOps and DevSecOps, then with just a few simple scripts on your Jenkins, you can work wonders using Llama & Ollama.

4. Enhancing Log File Analysis

One of my practices that I usually like to do is to analyze the application logs to search for hidden issues that can be fixed before it is too late, or even sometimes I find it hard to understand what is in the log file and what it is trying to tell. Once again, GenAI can come to help.

Check out my latest experiment using Llama & Ollama to build a simple log analysis.

5. Engage with Documents without Reading Them All

Visualize having a PDF file at your disposal — instead of sifting through every line, envision using OpenAI, LangChain, and Streamlit, as demonstrated in my experiment, to swiftly engage with your PDF content. Now, picture tech companies offering a service where developers can seamlessly navigate through documentation, enhancing accessibility and efficiency.

Chris Shayan on LinkedIn: I am getting to learn on how to use LangChain & AI4Finance Foundation…

6. Modernizing the CI/CD

In this video Matt Martz presents an implementation of a Slackbot that governs and executes deployments using Github actions. The Slackbot is augmented by GenAI (Bedrock) to conduct summaries of changes that pass through the CI/CD Pipeline.

With this bot, you are able to practice Continuous Deployment right from your Slack instance, as it is also able to summarize changes for you.

Conclusion

In conclusion, Generative AI (GenAI) is rapidly transforming the software development landscape, offering a plethora of benefits to improve Developer Experience (DevEx). As we’ve explored, GenAI can act as a powerful coding assistant, generate code snippets and test data, streamline log analysis, and even process documents without needing to read them entirely. These are just a few examples, and the possibilities are constantly expanding.

The potential of GenAI to automate mundane tasks, improve code quality, and empower developers to focus on creative problem-solving is undeniable. However, it’s important to remember that GenAI is a rapidly evolving field. To stay ahead of the curve and reap the maximum benefits, companies must be proactive in exploring and implementing these transformative tools. By embracing GenAI, organizations can empower their development teams, accelerate innovation, and gain a significant edge in the competitive tech landscape. The future of DevEx is undoubtedly intertwined with GenAI, and the time to act is now.