Anika Mantripragada is a designer at USC whose portfolio gets into the substance of working with AI.

She has adapted an existing design system for AI-generated output and built an iOS app to solve a problem she encountered herself. Those projects give her technical interests weight beyond the introduction.

Her portfolio also experiments with small interactive demonstrations that help explain the work. Let’s look at why those things stood out, and where clearer presentation and a more careful finish could help her get more from them.

Let’s have a deeper look into what makes it work so well!

Anika’s portfolio gave me a clear picture of someone engaging with how AI changes product design.

That begins in her introduction and continues through the GitHub activity she makes visible. But the projects do the important work of supporting those signals.

A contribution graph can show activity. Her case studies explain what she has been doing and why it matters.

During her Sauce Labs internship, Anika worked with an existing design system and adapted it to support more accurate AI-generated output.

This caught my attention because it overlaps with work I am doing myself. Making a design system useful to AI involves thinking about what information a model needs, how that context is presented, and how to determine whether the resulting output is good enough.

Anika goes beyond showing that she can enter a prompt and produce a screen.

She investigates the output, changes the context around the system, and tests what happens afterward. That testing matters because a model can produce different results across attempts. One convincing screen does not establish that the approach will hold up.

What she describes includes evaluation work, even though she does not foreground that term. I would consider naming it more clearly because it is a valuable part of the contribution.

Working this way during an internship is impressive. She took something that already existed inside a business and helped make it usable in a different kind of workflow.

That gives a hiring team a way to understand where her skills could apply within their own organization.

Cue provides another kind of evidence.

Anika wanted to recover saved TikToks when they became useful. Planning a trip to Oslo meant returning to saved cafes, restaurants, and guides, then trying to turn that collection into something she could act on.

I recognize the problem from my own saved Instagram Reels. Saving something is easy. Finding the useful piece again at the right moment can be much harder.

Anika built an iOS app to address it.

She thought through how it would work with TikTok, built the supporting functionality, and created an interface that uses familiar iOS conventions. The case study stays focused on a manageable problem rather than expanding into an enormous feature list.

I do not know whether Cue has reached the App Store. That does not prevent the project from demonstrating something useful.

It shows her taking a problem through design and implementation with AI, including the technical questions needed to make the experience work. Parts would need refinement before a broader release, but there is already substance to assess.

These two projects complement each other.

Sauce Labs shows Anika working within an existing business and design system. Cue shows her initiating a project, deciding what to build, and carrying it far enough to use.

For me, that combination says more than a collection of speculative app screens. It suggests she can investigate a difficult product problem through working prototypes and have a more informed conversation with engineers about what it would take to solve it.

Her portfolio contains smaller experiments too, but these two projects were enough to make me want to explore further.

Anika also experiments with how a case study can explain a result.

In the design-system project, she includes a small interactive comparison that lets the visitor examine generated output side by side. It helps communicate what changed as she improved the context given to the model.

That gives the interaction a purpose.

Someone can pause during the scroll, try the demonstration, and understand a point that might otherwise need more text or several static images.

I like how little commitment this asks from the visitor. They can explore one contained idea without opening a separate prototype or learning how to navigate an entire application.

The presentation still needs polish, which I will come back to. But the underlying choice is effective: let someone inspect the difference rather than asking them to accept a written description of it.

There are other places where this approach could help a case study.

A research chart might begin with its main finding visible, then let someone explore a comparison. A prototype could isolate one important behavior. A result could reveal the change between two states.

Anika is showing generated outcomes here, rather than an interactive research dataset. The broader opportunity is to make selected parts of a story easier to explore at different levels of detail.

That can preserve depth without placing all of it in front of the reader at once.

I would still use these demonstrations selectively.

A case study needs to communicate its central story while someone scrolls. Interactions can reward attention, but expecting a visitor to click through every section would create another kind of reading burden.

One or two useful moments may be enough to make someone stop and examine the work more closely. Repeating the same interaction throughout a page can make it feel like another task to complete.

Anika’s comparison works because it helps explain the contribution. That is the standard I would keep applying as she develops this approach.

The depth I found in Anika’s projects was not immediately apparent from the homepage.

Her introduction points toward working with AI and agents. Looking at the project previews, however, I could have assumed I was seeing a conventional software dashboard and a TikTok-related design concept.

Those first impressions undersell what she did.

One project involved transforming how an existing design system supports AI-generated output. Another involved designing and building an iOS app herself. I want those contributions to be visible as soon as I reach the work.

At the moment, product names and internship labels receive attention before the explanation of the contribution.

Sauce Labs, Cue, and TRIPP mean different things to someone who already knows them. A visitor encountering those names for the first time needs more help understanding why the work deserves a click.

Anika already has useful project descriptions. I would make them larger and visible by default, rather than depending on a hover state.

The homepage should carry the story forward from her introduction.

If she introduces herself as someone who designs and builds with AI, the next thing I see should help me connect that statement to a project. I should be able to understand what she changed or built before deciding whether to read the case study.

Cue’s default preview makes this harder because it shows TikTok rather than clearly introducing her own product.

TikTok is part of the context, but a visitor also needs to see what Anika added and what Cue enables. A short sequence could establish the source material and then move into her app.

The videos could also play automatically. With a small selection of case studies, I would consider that a reasonable way to bring more of the work into view without requiring another action.

Choosing the right opening frame and sequence matters as much as making it move.

I spent time inside the projects and found the connection. A hiring manager making a quick first pass may not get that far.

Anika has evidence for the profile she is presenting. I would bring that evidence onto the homepage and make it much harder to miss.

Anika’s work held my attention strongly enough that visual polish became a later consideration during the review.

It still needs attention.

The overall portfolio has a clear visual direction, but smaller inconsistencies appear across typography, screenshots, mockups, and video. Together, they make parts of the presentation feel less resolved than the work itself.

One video contains a noticeable flash between scenes. I initially wondered whether it was a browser issue, but it appears within the recording.

That is worth fixing because it interrupts the viewer at the moment the video should be helping explain the project.

The interactive demonstrations also need a more deliberate visual pass.

Some of the typography and layout feel like an initial generated implementation that has not yet been fully refined. I cannot know how Anika built each element, but the presentation should feel as considered as the explanation around it.

In one section, there are many competing text sizes and treatments. Labels and headings that seem to serve similar purposes receive different levels of emphasis.

I would simplify those styles and define a clearer hierarchy across the case study.

A section heading should have a consistent role. Supporting labels should help organize the content without competing with it. A different color or size should communicate a difference the reader needs to understand.

Image preparation needs a similar review.

I noticed screenshots with visible scrollbars, inconsistent corner treatments, and content cropped awkwardly inside mockups. Some experiment imagery also looks softer than the surrounding material.

Each instance may be small. Seeing several during one visit makes the finish harder to overlook.

I would review the published pages as a visitor, checking every asset at the size it is displayed. That should include the first frame of a video, its transitions, the edges of a screenshot, and the point where content meets a device frame.

A second pair of eyes would help too. When you have built and reviewed something yourself many times, it becomes easier to stop noticing a flaw you already know is there.

This does not change my assessment of Anika’s ability to design. It identifies another pass the portfolio needs.

For roles with a high visual bar, that finish may influence whether someone continues looking. Her technical work gives her a strong reason to earn their attention; the presentation needs to hold up alongside it.

Anika’s portfolio contains the kind of work I want to see more early-career designers showing.

Her Sauce Labs project demonstrates that she can engage with an existing design system, improve the context available to AI, and test the results. Cue shows her taking a problem from her own life and building a focused application around it.

The interactive comparison adds another useful dimension by letting visitors inspect part of the contribution themselves.

I would make those strengths clearer on the homepage and give the whole portfolio a careful visual review. The substance is already there, but I had to spend time uncovering some of what made it compelling.

These are projects I would want to discuss in an interview: how she tested the generated output, what she learned while building Cue, and which decisions changed once she had something working in front of her.

Still struggling to get your portfolio off the ground?

Don’t want to spend weeks learning yet another tool? Framer is my top recommendation for building your portfolio — fast, clean, and without the usual headaches.

If you’re just starting out (or even if you’re not), I think Framer is a perfect fit. Here’s why:

And that’s just scratching the surface. I wrote more about why I recommend Framer here—but honestly, the best way is to try it for yourself.

Affiliate disclaimer: I only recommend tools I personally believe in. Some links in this post are affiliate links, which means I may earn a small commission if you choose to purchase — at no extra cost to you.

Do you want your own portfolio reviewed in-depth with a 30-minute advice-packed video review? Or do you require mentoring to figure out a proper strategy for your job search?

I got you!