Zaria Parvez on Duolingo’s TikTok Voice and Building Creative Teams

Zaria Parvez, Duolingo’s senior global social media manager, joins Harry Stebbings to discuss the brand’s TikTok strategy, creative teams and the pressure of viral success. She describes treating short videos as episodes in an ongoing story, while measuring views and app discovery separately. Her results and interpretations are the company employee’s account as of the January 2024 interview.

Parvez graduated in 2020 after studying advertising and marketing. She chose Duolingo partly because its smaller-city setting appealed to her, and partly because its education mission and multicultural environment felt personal. For her first year, she says, her social posts were conventional brand messages.

When employees returned to the office, she noticed a large green owl costume sitting in a corner. Already a TikTok user, she made a funny video with the mascot. The first experiments were less provocative than later posts, but their unexpected, human quality suggested a different direction for the account.

TikTok was familiar to Parvez from college and the pandemic. She recalls the platform announcing a billion users in September 2021 and reasoning that time spent there competed with time in the Duolingo app. Her aim was to bring the lesson reminder into a space where people were already paying attention.

She considers TikTok a mix of art and platform knowledge. A creator needs to understand trends and audio, but also make a concise, entertaining piece. When hiring, she values the ability to take a concept, choose an appropriate sound, film it and write a caption more than a list of prestigious advertising employers.

Parvez suggests starting by adapting trends with a distinctive point of view and giving credit where appropriate. But copying formats only goes so far. As audiences become familiar with a brand, she believes original characters and continuing storylines give them a reason to return.

She likens Duolingo’s account to a sitcom: each video is an episode, some designed to go viral and others to build the story. Recurring characters include Duo, legal counsel Steve, Lily and Duo’s fictional infatuation with Dua Lipa. Collaborations and callbacks reward viewers who recognize the ongoing jokes.

Asked whether new followers will understand the backstory, Parvez says marketers should trust audiences to make connections and explore if they are entertained. She believes a brand can be silly on social media and still provide a serious learning product. That confidence is her view as a Duolingo employee, not an independent assessment of educational efficacy.

Not every video has to explain the app. Parvez says most people visit social platforms for a break or entertainment, not for language lessons from a brand. Her team tries to meet them in that context, while maintaining a link to the product elsewhere in its content or conversations.

Early ideas came from office jokes and the audience itself. People had long confused Duolingo with Dua Lipa, and “legal Steve” was already a workplace nickname. Parvez calls the comment section a creative brief: if viewers keep making a joke or asking for something, creators can develop it.

The social team had grown from Parvez alone to a small group with a head of social, coordinator and another manager, with more hires planned. Instead of relying solely on a meeting-room brainstorm, members share trends and ideas in Slack, film rough versions and gather feedback. Some experiments never get published.

Parvez doubts the same early creative process would have worked as well remotely: the costume and spontaneous office environment gave the team material. She says resourcefulness matters too, such as using a water bottle as a pretend microphone rather than waiting for professional equipment. This is her experience, not a universal remote-work verdict.

As the account grew, she felt pressure from the high standard it had set, especially as a junior employee in her first job. She is still learning management and workplace skills alongside the public success. An underperforming video, she says, does not mean the team has failed.

TikTok, Instagram Reels and YouTube Shorts do not always reward the same material. At the time, Parvez estimates about 75% of the team’s content was reused across platforms and the rest adapted. Colleagues familiar with different channels and audiences helped adjust the same idea for each context.

She does not claim a reliable viral button: platform behavior changes, and repeated editing can make a creator lose perspective. Her test is whether she personally finds the video entertaining enough to share in a group chat. She also asks her teenage nephew for candid feedback from outside the brand.

Parvez values quality over posting merely to satisfy a frequency rule. She has taken breaks when burned out and found the audience continued with its life; after returning, the team could create successful videos again. That experience helped her take the work less personally, though it is not a guaranteed posting strategy for every account.

A brand’s channel choices should follow its own strategy and audience. Duolingo’s guiding idea is that language learning is hard, so the brand makes it fun; another is that Duo appears everywhere to remind people to practice. That makes a presence across several platforms sensible for her team, but a narrowly targeted business may choose differently.

With millions of followers, Parvez has learned to consider the reach of a post or comment before sending it. She has crossed a line in the past and received feedback. She tries to balance the responsibility of a large audience with an environment that permits experimentation and learning from mistakes.

She recalls a My Little Pony trend whose execution looked more sexual to others than she intended. The CEO asked the team to remove it even though it was performing well. The lesson, she says, was to communicate why a risky idea works to people with different experiences and media habits.

Parvez describes direct feedback from the CEO as a conversation rather than a battle. In some cases she has proposed changing an audio track or text before reposting. She says social, legal and leadership all want the account to succeed, even when they disagree about where the boundary lies.

The account began as a brand-awareness experiment. Working remotely with limited resources, Parvez first commented on other videos; once those comments drew attention, original videos gave visitors something to find on the profile. At Duolingo, one million views became an internal threshold for calling a video viral.

For her, cultural impact is broader than a view count: people mention the account in group chats or cite it when applying to work at Duolingo. The company also added TikTok to its “how did you hear about us” survey and observes increases in reported discovery around viral posts. Timing and self-reporting suggest a relationship but do not establish precise causal conversion.

Stebbings asks whether followers and likes become paying users. Parvez counters that conventional social posts also deserve scrutiny and says her team can experiment inexpensively. She estimates roughly 95% of its impressions were organic at the time; that is her company-specific estimate, not a measured return on spend or a subscriber conversion rate.

Stebbings says his team spends $25,000 a month on short-form content across channels. Parvez prefers tight constraints and believes scarce resources can prompt creative solutions. Her preference does not mean production is free: staff time, tools and other resources still matter.

She is skeptical of outsourcing the core brand voice to an agency. Duolingo tried an agency when it began on TikTok, she recalls, but the content gained traction after an internal team took over. She believes employees immersed in the mission can respond with a closeness and attention that are difficult to reproduce across multiple agency clients.

Asked about internal reactions to her personal visibility, Parvez credits an early manager, Michaela, for publicly recognizing her work. She describes friction more as the growing pains of a brand becoming social-first than as jealousy. The support of leadership made it possible for a junior creator to take responsibility and receive recognition.

Parvez credits both opportunity and work: supportive family, education, internships and a brand willing to take risks created access. Her “20% luck, 80% work” framing is a personal shorthand, not a measured breakdown. She is proud that she acted on those opportunities and believes others given similar access can succeed too.

For creative hiring she recommends people willing to question conventions and push boundaries, whom she calls troublemakers. She recalls a nontraditional résumé and unusual interview questions, as well as having to persuade managers to try ideas. She values people who can manage the tension and learn when an experiment creates problems.

She also sees strength in people who have navigated intersecting cultural backgrounds. Growing up in a Pakistani household in a predominantly white American city, she learned to balance different expectations. Her broader claim about such backgrounds is a personal perspective, not a rule about every candidate.

Bold social work can create friction across teams. Parvez recalls colleagues objecting to a post because it intersected with a children’s product. Her approach is to discuss the concern and explain the social context; she sees respectful disagreement and a thick skin as part of working across functions.

The team sometimes removes a particularly poor post, but usually leaves underperforming work up. Parvez says she was experiencing a difficult period for content at the time and was trying not to equate the green mascot with her entire identity. Stepping back helps her accept that not every video will be a hit.

Cultural moments often outperform scheduled content in her experience. The team keeps a loose calendar and gives members a veto when a timely idea needs to go out immediately. Plans rarely extend much beyond two weeks, and creators are expected to postpone their own planned posts when a more relevant opportunity emerges.

For Parvez, one of the hardest parts of content work is separating personal attachment from critique. A team that has succeeded can still struggle on a new platform or go through a dry spell. She sees adapting and taking feedback as essential, even when the process feels like starting over.

Her shorthand for balancing entertainment and product is “candy to the medicine.” The entertaining video attracts attention; a caption or comment can then mention language learning or a premium feature. She gives a Dua Lipa joke that included a Duolingo product reference as an example, not a promise that every viewer converts.

She would rather give responsibility to someone who genuinely enjoys TikTok than require a conventional marketing background. Her ideal hire might even be a comedian: she thinks marketing skills can be taught, while instinct for entertaining video can come from outside the profession. This is her hiring preference.

Parvez admires accounts including Scrub Daddy and the New Jersey state government, and cites Liquid Death’s longer-form work. She says she values people who make strong creative work more than having a single ideal brand. Duolingo has also collaborated with Scrub Daddy.

She thinks standing out on social platforms has become harder as more brands participate. At the time she worried about a possible TikTok ban and about sustaining storylines beyond individual trends. These were concerns in January 2024, not a statement about the platform’s later legal status.

The team had not found a direct correlation between a viral TikTok and growth on other social channels, she says. It had observed a possible relationship with app downloads, while earlier reposts sometimes attracted followers on Twitter. She does not provide a quantified causal estimate.

Looking ten years ahead, Parvez says she wants a stable job, good work-life balance and a life that is not defined by work. She does not currently aspire to be a chief marketing officer simply for the title, though she recognizes her ambitions may change. Her account closes on the value she places on being content outside a career treadmill.