Uplift AI: Giving a Voice to the 42% Who Can’t Read
Uplift AI is a mission-driven technology startup founded in 2024, part of Y Combinator’s Summer 2025 batch. The company is building foundational speech understanding and generation models for the five major languages spoken in Pakistan—Urdu, Punjabi, Pashto, Sindhi, and Balochi—languages that have been largely ignored by Big Tech despite their millions of native speakers.
Pakistan, the fifth most populous country in the world, with over 240 million people and a GDP of $411 billion, faces a startling problem: 42% of its adult population cannot read in any language. In a digital world still largely reliant on text-based interfaces, this becomes a massive barrier to inclusion. Uplift AI is tackling this head-on with voice interfaces tailored to the linguistic realities of its population.
From banking and healthcare to agriculture and government services, Uplift AI’s APIs and SDKs help businesses deploy powerful voice solutions—finally enabling people, even those with no literacy, to access critical digital services.
How Did Uplift AI Begin?
The origins of Uplift AI stem from curiosity, frustration, and a commitment to impact. Hammad Malik, co-founder and CEO, left Apple in November 2023 with no set plans but three very different ideas in mind: improving political transparency in the U.S., building luxury tiny vacation homes, and developing voice interfaces to make knowledge more accessible.
Of these, the third idea was meant to be just a side project—a nonprofit, even. But within a week, Hammad found himself unable to step away from it. The other ideas were abandoned. The deeper he delved into the academic and technical landscape of speech technology, the clearer the gap became.
While English voice AI had seen rapid advancements, most other languages were left far behind. The reasons weren’t rooted in data scarcity alone—a common misconception—but in systemic neglect and suboptimal engineering decisions at nearly every layer of the tech stack. Academia didn’t incentivize innovation in “small” languages. Big Tech’s organizational structures made it almost impossible to do the hard, integrated work required for high-quality voice models.
After a research binge of 30+ papers and months of prototype building, a visit to Pakistan changed everything. Hammad spoke to users firsthand—farmers, health workers, everyday citizens—and saw the scale of the issue up close. He realized that entire economies were being held back simply because people couldn’t interact with digital tools built only for the literate.
What began as a side project turned into a full-fledged startup: Uplift AI.
Who Are the Founders Behind Uplift AI?
The founding team behind Uplift AI brings a rare mix of world-class experience and grassroots motivation. They’ve worked on some of the most advanced voice technologies in the world—including Amazon Alexa and Apple Siri—and now they’re redirecting their expertise to regions and languages overlooked by global tech giants.
- Hammad Malik, CEO, previously worked on Siri at Apple and led projects at Amazon Alexa. He studied at Cornell and LUMS (Lahore University of Management Sciences).
- Muhammad Bin Sabir, COO, is an entrepreneur and LUMS graduate who previously founded Alt Ventures, an accelerator focused on Pakistani startups.
- Zaid Qureshi, CTO, worked on cutting-edge infrastructure at AWS Bedrock and Amazon Supply Chain, bringing a deep understanding of scale and systems.
Together, this team brings not only technical capability but also cultural insight and a proven ability to ship high-impact software.
What Problem Is Uplift AI Solving?
Digital exclusion is often talked about in terms of internet access or smartphone penetration. But there’s another, less discussed dimension: linguistic exclusion.
In countries like Pakistan, where nearly half the adult population cannot read, text-based digital interfaces are fundamentally inaccessible. Think of an illiterate farmer trying to access government subsidy information. Or a rural mother seeking telehealth advice. If the interface assumes literacy in English or even Urdu, it fails before it starts.
Moreover, most so-called “voice assistants” today—Google Assistant, Siri, Alexa—either don’t support local languages at all or do so with limited accuracy and poor expression. That’s because the foundational models were not built with these languages in mind. From tokenization and phoneme mapping to pronunciation and emotional inflection, everything is optimized for English or other dominant tongues.
Uplift AI builds from the ground up. They gather their own data, record professional voice samples in local languages, and train their own models. By owning the full stack—from data collection to model deployment—they can optimize every layer for quality and contextual relevance.
How Is Uplift AI Different from Big Tech?
Unlike the sprawling and siloed structures of Big Tech companies, Uplift AI operates with end-to-end control. When a team needs to collect voice data, they don’t need legal, PR, and contracts to approve it—they just do it. When they want to tune tokenizers or develop expressive TTS (Text-to-Speech) in Urdu, there’s no need to fight for another team’s attention on a multi-quarter roadmap.
This autonomy allows Uplift AI to focus relentlessly on quality. Their philosophy is simple: "You get the models you earn." And in underserved languages, the only way to earn them is to do everything yourself, properly.
Where Big Tech optimizes for scale and generalization, Uplift optimizes for nuance, locality, and inclusion. Their systems aren’t just functional—they’re lovable, expressive, and culturally fluent.
What Are the Use Cases and Impact?
Uplift AI’s technology is already finding traction across multiple domains:
- Banking: Enabling voice-driven account access, bill payments, and balance checks without needing to read a single word.
- Agriculture: Helping farmers get audio-based guidance on crop health, weather, and pricing in their native tongue.
- Healthcare: Allowing voice booking of appointments, symptom checks, and medication instructions—crucial for patients in rural areas.
- Government Services: Facilitating access to public records, social welfare, and voter information through simple spoken commands.
These are not just conveniences—they’re lifelines. For many users, it’s their first-ever experience of digital empowerment.
What’s Next for Uplift AI?
The startup’s roadmap is ambitious. Uplift AI plans to expand its language models further, deepen integrations with enterprise partners, and grow its footprint across South Asia and other linguistically rich, digitally underserved regions.
The team is also investing in research partnerships and open-source contributions to elevate the entire field of multilingual voice AI. Their goal is not just to dominate a niche but to redefine what inclusive technology means globally.
They are actively hiring engineers, researchers, and linguists who believe in building tech that uplifts, not excludes.
Why Does This Matter Globally?
Uplift AI isn’t just a Pakistani success story in the making—it’s a proof-of-concept for a much broader truth: the next billion users will not speak English, and they won’t all be able to read.
Voice is the future of interfaces, especially in mobile-first regions. But only if that voice speaks their language—with fluency, context, and dignity.
By building foundational voice models for overlooked populations, Uplift AI is showing that inclusion doesn’t require compromise. It just requires care, expertise, and a different set of priorities.
As the digital world expands, the question isn’t whether underserved languages will be supported. It’s who will do it right, and for whom.
Uplift AI has chosen its audience: the unheard, the overlooked, and the next billion. And they’re giving them a voice.