Tanul Singh

I come from a very small town in Uttar Pradesh, India. Growing up, the only path to success anyone knew was: study hard, clear IIT, get a job at an MNC. I studied hard through my 12th, but I didn't clear IIT. I ended up at the best college in my state, studying Mechanical Engineering — a branch I got because of a bad rank.

But my heart was always in Computer Science. In my second year (2016), a senior mentioned Machine Learning. I found Andrew Ng's course, tried the digit recognizer, and was instantly hooked. My father was a maths teacher — seeing how we can express the intuitive things we experience in life through mathematics just amazed me.

I lost my father in my second year of college. With no earning member in the family, everything fell on me. But I was so hooked on ML that I didn't stop. I went to college during the day, tutored kids to raise funds, and studied Python and ML deep into the night.

Being in India, I didn't have access to good teachers or people who were doing ML work at the time. But I found Kaggle — I learned from notebooks shared by the best in the world, people I'd never meet. After two years of this routine, I became a Kaggle Competitions Expert and secured a Data Science job — the first person from my college to do so.

In September 2020, I made a promise to myself: someday, I will work as a Research Scientist at a world-class lab, leading the frontier of AI. Every step since then has been toward that goal — Kaggle Grandmaster, Stanford courses, published research, a US patent, building production ML at LevelAI, and now working at Apple where I collaborate with the ML Research team.

My work at Apple sits at the intersection of research and applied AI — roughly half my time goes into ML research problems and the other half into building production systems that ship to millions of users.

On the research side, I collaborate with Apple's ML Research team to investigate open problems in LLM generalization, knowledge representation, and model robustness. On the applied side, I build and deploy large-scale NLP and generative AI systems — from multi-agentic architectures to safety and guardrail frameworks.

It's the kind of role I'd been working toward since 2020 — one foot in research, one foot in production, and both feet in problems that actually matter at scale.