Inderjit Singh Chahal Machine Learning Engineer
— J. J. Sylvester, 1864
I take that conviction to work — turning rigorous math into production deep-learning systems: autonomous medical imaging, custom OCR objectives, attention-based NLP. Models that ship.
01 A custom objective using Fourier-transform loss that makes reliable predictions even for single-character maps — pushing detection quality past the standard baseline on a public benchmark.
02 A robotic, intelligent radiologist — a fully automated device that tests and quantifies the functions of a radiologist using state-of-the-art AI and mathematical tools, delivering near-human and, under some conditions, better-than-human results.
03 An end-to-end package to monitor and secure ATMs. A custom-trained AI watches every machine without human assistance — reading people count, posture, suspicious objects and aggressive expressions — and raises alerts on customizable conditions, while all streams feed a central GUI for single-point monitoring.
04 An encoder–decoder architecture with trainable embeddings and Bahdanau + Luong attention — capturing intent and long-range dependencies by focusing on the right words and retaining far more context than a vanilla encoder–decoder.
I like problems that are new, different and rare in how they're solved — and I'll spend hours optimizing code or collapsing a complex problem into a simple implementation.
I'm a Google Certified Professional Machine Learning Engineer and Google Certified Cloud Engineer, and I love designing architectures for cloud-based applications. I enjoy debugging the hard, tangled bugs, and I give back on community platforms like Stack Overflow and Reddit.
Away from the day job, I work with financial markets — building automated strategies augmented by AI, where probability and risk modelling earn their keep.
▶ Watch the 2-minute intro
