I have always been drawn to the places where meaning is not immediately clear.

My work begins with language: with poems, silences, mistranslations, political signals, machine outputs, and the many human things that resist easy interpretation. As a scholar of Persian, Urdu, Arabic, and Indo-Persian literary traditions, I was trained to read slowly — to pay attention to metaphor, contradiction, longing, excess, and the meanings that gather between words. That training has shaped everything I do.

Whether I am working at a policy think tank, analyzing geopolitical sentiment, or doing AI Trust & Safety work at Meta, I return to the same practice: reading what is difficult, staying with ambiguity, and turning uncertainty into insight. At FilterLabs AI, this meant working with messy, multilingual, cross-cultural data and asking what headlines or surface-level metrics could not fully tell us: what forms of sentiment were moving beneath the visible narrative, what signals were emerging before they had been fully captured by public discourse, and how fear, exhaustion, anger, resignation, or the desire for political change appeared across fragmented digital spaces. In one project, our work examined whether Iranians were expressing support for regime change, with findings later published in The New York Times.

In AI Trust & Safety, this same attention to ambiguity becomes a way of thinking about governance, harm, and human consequence. At Meta, I worked with high-risk and edge cases across domains such as hate speech, terrorism and violent extremism, child safety, self-harm, privacy, and other sensitive areas where language, culture, safety, and model behavior intersect. I am interested in what happens when technological systems try to interpret human language, risk, and context — and what those decisions mean for the communities that have to live with them. My work asks where policy misses complexity, where harm remains hidden because it does not fit neatly into a category, and how governance can become more attentive without flattening people into data points, labels, or risk scores.

Alongside this work, I write fiction and poetry in Urdu, English, and Persian. Writing is one of the ways I have learned to think: to notice texture, consequence, contradiction, and feeling — the parts of human life that are often hardest to measure, but most important to understand. Across scholarship, public writing, geopolitical analysis, and technology policy, I keep returning to the same questions: How do we read what is not yet fully legible? How do we make sense of ambiguity without erasing it? And how do we build systems that remain answerable to human life?

I have presented my work at more than fifteen international conferences, published peer-reviewed academic articles and op-eds, and published poetry and fiction in Urdu, English, and Persian.