I am a third-year Computer Engineering student at Kantipur Engineering College, Tribhuvan University, Kathmandu, Nepal. I work at the intersection of LLMs, multi-agent systems, and vision-language models. I'm part of the Fusemachines AI Fellowship (AIF 2026) and the Leapfrog Student Partnership Program (LSPP 5th cohort).
Somewhere beneath the engineering, I'm genuinely curious about something harder to pin down: whether language models can have something like thought. Not agents calling tools, not chain-of-thought prompting, but actual internal reasoning. I don't have answers yet. That's kind of the point.
Outside of my academic coursework, I also:
A practical, intuitive exploration of SimCLR, MoCo, BYOL, Barlow Twins, and CLIP, detailing what actually matters when building representation learning systems.
A mathematical deep-dive on temperature in LLMs, logits, softmax, and Hinton's concept of dark knowledge, revealing why creativity is not just randomness.
A reflection on Searle's Chinese Room argument, the slippery nature of human comprehension, and whether large language models are truly just syntax without semantics.
A reflection on Hume's skepticism, Judea Pearl's ladder of causation, and why causal reasoning is machine learning's last and hardest ceiling.
A periodic, focused newsletter on robust AI architecture and software systems. Straightforward engineering with zero fluff.