Who Keeps Your Machine Learning Work Updated?
Machine learning portfolios need clear ownership for updates, access, and follow-up—here’s how to decide who handles them after launch.

A machine learning portfolio like Ashley Miller’s Data Science Engineer Portfolio shows real work: live models, streaming data pipelines, and case studies. But after launch, someone must keep it working. Broken links, outdated case studies, or missing project details can make the portfolio look untrustworthy. The question isn’t if updates will be needed—it’s who will handle them. Explore Data Science Engineer Portfolio
This portfolio template includes machine learning projects, distributed data architectures, and research articles. It’s designed for engineers who build production systems like fraud detection pipelines or real-time feature stores. But the template itself doesn’t assign owners for updates, access, or follow-up. That’s a choice you’ll need to make.
For example, Ashley’s portfolio highlights a fraud detection system processing 25,000 transactions per second. If a client asks about the model’s accuracy six months later, who answers? If a new team member needs access to update a case study, who sets it up? These aren’t technical questions—they’re about who is responsible.
Three key tasks that need an owner
1. Fixing broken links or outdated details. Machine learning projects often rely on external links to papers, datasets, or tools. If a link breaks or a project description becomes outdated, someone must update it. In Ashley’s portfolio, a case study on a streaming feature store includes benchmarks like ‘1.8B events/day.’ If those numbers change, the portfolio should too.
2. Adding new projects or research. A portfolio isn’t static. New models, improved architectures, or published articles should be added over time. For instance, Ashley’s portfolio includes an ‘Engineering Note’ on sub-millisecond feature stores. If she publishes another note, who adds it to the site?
3. Answering client or hiring manager questions. A portfolio isn’t just for show—it’s often the first place clients or recruiters go to evaluate your work. If someone asks about the latency of a real-time model (like the ‘sub-15ms’ fraud detection system), who provides the answer? If the portfolio doesn’t reflect the latest work, it sends the wrong message.
Who should own these tasks?
The answer depends on your team’s structure. For a solo data scientist, the owner is obvious—you. But for teams, it’s trickier. Some options include: the data science lead, a dedicated technical writer, or even an external contractor. The key is clarity: assign one person (or a small group) to handle updates, access, and follow-up.
For example, if your team builds distributed streaming architectures like Ashley’s, the data science lead might own updates to technical case studies. But if your portfolio also includes research articles, a technical writer or editor might handle those. The goal is to avoid ‘update rot,’ where no one takes responsibility.
A limitation: No built-in access controls or reminders
This portfolio template doesn’t include tools to manage user access or send reminders for updates. That means you’ll need to set up those systems separately. For example, you might use a shared calendar to track when case studies need reviewing or a project management tool to assign update tasks. These aren’t part of the template—they’re requirements you’ll need to discuss with TechiesAIE. Explore TechiesAIE software services
If your team works remotely or across time zones, this becomes even more important. For instance, Ashley’s portfolio includes projects built with tools like PyTorch, Spark, and Kafka. If a team member in another country needs to update a case study about a Spark pipeline, they’ll need clear access rules. The template doesn’t handle this—you’ll need to plan for it.
What a similar project would need
If you’re commissioning a custom machine learning portfolio, here’s what you’ll need to discuss upfront: 1) Ownership: Who will handle updates, access, and follow-up? 2) Access rules: How will you control who can edit the portfolio? 3) Reminders: Will the system alert owners when updates are needed? These aren’t features of the template—they’re decisions you’ll make during the build.
For example, if your team builds production ML systems like Ashley’s, you might want the portfolio to include a ‘last updated’ timestamp for each case study. This shows clients the work is current. Or, if your portfolio includes research articles, you might need a way to flag ‘draft’ or ‘published’ status. These are custom requirements, not part of the template.
TechiesAIE can adapt this portfolio template to fit your needs. For instance, we’ve built similar tools for Software Engineer Portfolios where ownership and access are critical. Or, if you need a custom solution for tracking ML projects, we can explore building something tailored to your team’s workflow.
The template is a starting point. The real work is deciding who will keep it running—and planning for the tools they’ll need to do the job.