Ada does not have a preference whether students work on a specific intern project or they are integrated into the team to work on tickets, so long as they are doing software engineering work that will help prepare them to be a hireable junior developer. Intern’s projects or tasks must be solidly based in coding or development, at least 75%.
Consider the SDE1 guidelines of your organization, and work towards those. The expectation should not be that interns come in meeting those expectations on day 1, or even by month 2-3. There is a reason these are 5 month long internships!
The following information will help you:
Scope an appropriate intern project and workload.
Choose a mentor.
Set up adequate support for the intern - we advise to start with a higher level of support, then ramping down to more independent work.
Intern projects are often unrelated to real world projects, and a new intern does not have the experience to identify that. The goal of the internship is not to rewrite a previous intern’s code or spend time on work that does not translate to the real world.
Writing tests is the task of whoever wrote the original code. If tests weren’t written when the code was written, is that code written in such a way that it’s conducive to testing? If tests weren’t written when the code was merged, how was that code approved and merged by your team? Would you make a full-time SDE write tests for someone else’s code?
Interns are here to learn how to be software engineers, not how to be program managers. Administrative work is disproportionately assigned to women and underrepresented minorities.
Code review is one way that we learn to write better code. If your intern is working on a long-lived branch that isn’t being regularly reviewed by the team and integrated to the mainline, they are missing out on critical feedback from seasoned engineers. Interns are here to learn how to do the job of a software engineer, and this job involves code review.
Team: [(e.g. Product name team)]
Ada SDE Intern: [(e.g. Intern Name, alias)]
Manager: [(e.g. Employee Name, alias)]
Mentor: [(e.g. Employee Name, alias)]
Onboarding buddy: [(e.g. Employee Name, alias)]
Internship: [(e.g. Fall 2021)]
Description of Project:
What does successful completion of this project look like?
What are the specific growth areas/skills that will be explored in this work?
What areas, if any, will be alterable during the internship?
Guiding Questions
Can you explain the proposed work to an SDE2 in two sentences or less, and they feel they could reasonably complete the work in two weeks without any additional information? This is a good indication of a right-sized three month project for an intern. Remember that if your intern completes a first project, they can work on stretch goals or start work on a second project.
Has this project ever been assigned to a previous intern? You cannot re-assign a project that was already assigned once and not completed.
Is this project an extension of another intern's project? You cannot assign a project that builds off of the work of a previous intern. If a previous intern's project is incomplete or needs additional features, assign that project to an SDE1 on your team.
If this work expands on existing code, are testing frameworks properly configured and conducive to writing tests? Is the existing code sufficiently tested? You cannot assign interns to write tests for existing code.
Do you own the codebase(s) that this project involves? You cannot assign an intern to do away team work in someone else's code base.
Does this project rely on APIs or data sources that your team does not own? If yes, do you have the necessary relationships and permissions (AAA, etc) for those APIs or data sources? If not, who on your team will be responsible for establishing those relationships and when will that work be complete?
What language(s) will the project use? [e.g. Java, Scala, TypeScript]
How many developers on the team have recent or current experience with that language?
What tools and resources will the intern need to learn in order to complete their project? [e.g. DynamoDB, Apollo, SQS]
Does the rest of the team think that this project is important? If the rest of the team does not believe this work is important, they will typically not provide the level of support that these interns need to be successful.
Background about the space and reasons for choosing this project, it should answer the following questions: Who are you as a team, what is it that you build or provide? Why is this a problem, and who is impacted? What is the benefit of solving this problem / why did you pick this project?
Provide a list of two to three ramp up tasks. These ramp up tasks should be relevant to the assigned project. Do not assign ramp up tasks in a codebase or service that does not pertain to the assigned project; setting up development environments is expensive and low-value, and context switching is particularly expensive to interns. Sample ramp up tasks could include: demonstrating that the intern has set up their development environment correctly and can run unit tests inside their IDE; successfully deploying code from origin/mainline to a test account (if that process is part of how your team tests code); making a small and unambiguous code change and publishing a code review.
Your intern is not yet an SDE1. It is not appropriate to assign an intern the task of designing software. Provide a link to an existing design document, or transclude that document here. If a design document is not yet prepared, list who is responsible for writing that document and getting approval from relevant stakeholders, and provide a date by which the document will be complete. If this work involves front-end work, provide UX mock-ups here as well.
How will you know that the project is completed successfully? These can be user stories, or they can be high level bullet points of the functionality.
Interns are learning how to do the job of a software engineer. They have very limited experience breaking down work into appropriately-sized milestones; it is your job to teach them how to break down their work. We recommend thinking about a milestone being a single code review. A code review should be the smallest possible testable unit of work. From the Intern Project Guidelines, "someone on your team should know exactly how to do what the intern is being asked to do." Ask them to write down the titles of the first five code reviews that they would expect the intern to publish. Limiting this list to five milestones helps make the list less imposing to your intern and allows you to adjust the size and ambiguity of milestones as your intern gets further into their project.
*You cannot assign a project that depends on integration with a) services that are in the process of being built, b) services that are owned by a team whose planning and roadmap you cannot directly influence.
Please provide links to all packages that will be modified by the intern. Please provide links to all services, APIs and data sources that this project will interact with.*
Detail the contact information and role of all stakeholders in the delivery of this project. Who will be using it? Who has a vested interest in its success?
The stated goal of an Ada internship is to produce a “hireable junior developer.” Make that goal your guiding principle.
Part of what interns are here to learn is how to do this job. Writing code for school projects and doing the job of a software engineer are barely related. Find the most compassionate, empathetic teacher on your team, and have the intern sit with and observe that person nearly full time for two weeks. Ask the mentor to do a lot of talking — talking about what exactly they’re doing, about what they’re stuck on, about how they make progress towards getting unstuck. Make sure your intern understands that their job during this period is to ask tons of questions about anything and everything, including stuff they might think they should already know.
This two week period should be enough time for the intern to watch a full-time engineer take a sizeable task through the implementation, test, code review, deployment, and completion cycle. Having a dedicated observation period helps build a strong foundation and helps develop your intern’s trust in their mentor.
Ada interns don’t have enough experience to have a preference, and some languages (like Python) are notoriously difficult to use within your company’s ecosystem. If your team primarily writes Java and TypeScript but your intern is working in Python, offering help to the intern will be more expensive to your team and therefore less likely to happen. Just tell them what language to use and provide them resources to ramp up on that language.
Setting up development environments is a time-consuming and low-learning-value task. It’s also a place where interns tend to get stuck for long periods of time. Identify the service that the intern will be working on and make sure that they are well configured for working on that service. If you move the intern around from one service to another between every task, expect every task to take ten times as long as you think it should.
If your proposed project includes, “For this project, we will integrate with an API that is scheduled to be complete by the end of Q3,” something is wrong. We know that in software things rarely go according to schedule. If your intern’s success is closely coupled with someone else’s ability to deliver, you’re setting your intern up for failure.
Other teams don’t care about intern projects. Interns have very little influence, and their requests often get shoved to the bottom of the stack. If you need to integrate with another team, have your intern pair with an experienced engineer so that they can see how the process works; don’t ask your intern to do it on their own.
If you can explain a proposed piece of work to an SDE2 in two sentences, and they could complete that piece of work in two weeks without any additional information, you’ve found a reasonable two month intern project.
As engineers, we should always be looking for the smallest testable logical unit of work. Finding a correctly-sized unit of work is a skill that takes time to develop. It’s up to you and your team to teach this skill. Here’s an example: Imagine a task to create a Lambda function that can read from DynamoDB. Without guidance, your intern is going to try to do that in a single code review and they are going to be completely overwhelmed. One small, testable, logical unit of work I see is standing up the DynamoDB table. You could absolutely create a code review for just the CFN/CDK code to create the table. Less than 20 lines of code, tested with a SAM/CDK deployment. Show them that that’s the size task they should be looking for and working on.
“Watch one, do one, teach one” is a common education refrain for a reason. Repetitive exposure is critical to deep understanding. Imagine your intern on an ever-increasing spiral — returning to the same concepts repeatedly, but each time slightly more complicated than the last.
An intern project is not the right time for your team to say, “Oh, we’ve always been meaning to try the CDK instead of writing raw CFN. Let’s have the intern figure that out.” Someone should be so familiar with the intern’s task that they have the entirety of the code pre-written in their head; that they could glance at an error message and know the solution to the problem; that they could scan a code review in under a minute and know that the intern is on the right track.
Make sure that your team understands that their success is tied to your intern’s success
Everyone on your team needs to be bought-in to the task of developing your intern into a hire-able junior developer. Everyone on your team needs to be invested in offering help. The environment we’re looking for is one where your team members are in competition with each other to provide the most proactive, comprehensive, and plain-spoken support. Reward that behavior.
It’s tempting to assign an internship project that’s just a “nice to have,” but if the rest of the team doesn’t think that the intern’s project is important, they won’t be invested in helping. Make sure that whatever the intern works on, the team actually wants that work done.
The only requirements are that the intern have the opportunity to learn and to demonstrate that they raise the bar as an SDE1. If you’re struggling to find an internship project, maybe that concept isn’t right for you. Many successful interns demonstrated that they raise the bar by doing a series of small 2-3 week projects more in line with what a typical full time SDE1 would work on.
Often, we see teams tell interns, “Just ask if you have questions!” Ask who? Where? How often or after how long? What’s the SLA for addressing intern questions on your team? Should they escalate to you if they don’t get help within that time frame? Provide them with a how-to-get-help framework.
Being stuck and figuring out how to get unstuck is the nature of this job; it’s what all engineers are doing all day long. But being stuck on the same problem without progress for long periods of time is not valuable. When you feel stuck, here’s what I want you to do:
Keep a log of what you are doing. For instance: what commands did you run, what did you expect to have happen, and what actually happened?
Timebox the problem. If you’re stuck on the same error for 20 minutes, it’s time to ask for help. If you make it past an error and are on to a new error, reset the clock to 20 minutes.
When you ask for help, please ask for help publicly when possible. I know that asking for help can be nerve wracking as a beginner, but our entire team is here to help you learn. Asking for help where I can see the ask is helpful to me me because:
I can see what kinds of things you’re getting stuck on and offer additional resources when I see a pattern
I can see how quickly you get a response and whether that response meets my bar for assisting an intern
Other people can benefit from your question and the answer
When you ask for help, please try to include the following things:
What task are you trying to accomplish? What thing are you doing that you’ve gotten stuck on?
What have you tried so far? What did you expect would happen when you tried that, and what happened instead?
What other resources have you reached for? Sage? Wiki? Onboarding docs? AWS documentation? How have those informed your attempts to unblock yourself?
When all else fails, you can always say “I am stuck and I don’t know enough to formulate a question” and we will find someone to pair with you until the problem is in a state that we think you can handle again on your own.
If you ask a question and don’t get help within the hour, please reach out and let me know so that I can find help for you.
Ada interns are so new to this field that they often don’t even know what to have questions about or how to formulate those questions. Asking for help in a remote environment can feel particularly daunting. You or a technical mentor may need to draw the question out of them by asking them questions first. Here some ideas for drawing out questions from your intern:
How are you doing today?
Can you walk me through what you’re working on this morning?
How are you feeling about this task?
On a scale of 1 to 10, how strong do you feel your comprehension is on this task? What parts of it feel like a 10? What parts of it feel like a 1?
Would you like someone to pair with you for a bit while you get started on this new task?
Is there anything about what you’re working on that you feel daunted by? Would talking about it with someone help you break that work down into more manageable components?
Special Thanks: Alyssa Hursh, a graduate of Ada Developers Academy, for originally drafting much of this content.