Fullstack Developer · AI · Machine Learning

Bhuwan
Regmi

I build software to solve problems I actually run into, and lately problems other people run into too. Sometimes that means a fitness app, because logging workouts by hand was annoying and I wanted to do it by voice instead. Sometimes it means an operations platform for a processing plant, built from scratch and now under review for company wide rollout. Right now it means a referral toolkit for community workers who sit across from someone in a bad moment and need to know what to do next. Live products with real users, and two bigger ones still in motion.

Things I've
built.

What I've
studied.

I'm a fullstack developer with a focus on machine learning and artificial intelligence. My certifications are in supervised and unsupervised learning, reinforcement learning, linear algebra for machine learning, and generative AI. They are all listed above if you want the detail.

What makes my work different is context. Everything I build comes from a real gap I watched play out in front of me, not a classroom exercise, and most of it gets built in the evenings after a full day of work.

EVP Connect, a full employee communication platform, started with a communication gap on a plant floor. Information was scattered across wall postings, individual emails, and paper forms with no unified system connecting it all. I built the entire platform solo on my own time and presented it to the executive team. IT approved it, and I designed and built a set of admin console features on the back of their feedback. It is now under review for company wide rollout. My vendor data analysis identified $2.71M in actionable savings. My AI apps serve real users today.

Living Links is the newest one, and the first I am not building alone. It is a referral toolkit for frontline community workers at a nonprofit in Nova Scotia, and it has taught me the part of this work that no certificate covers. Taking a rough idea out of a conversation and turning it into something people can actually use. Deciding what not to build. Staying close enough to the people using it that what they ask for lands in the product within a day or two.

I use AI tooling heavily and I do not pretend otherwise. It is fast. What it does not do is decide what should be built, understand who gets hurt when it goes wrong, or notice that the wording on a screen is quietly misleading someone. That part is still the job, and it is the part I care about.

I learn fast, I ship faster, and when I get something wrong I say so and fix it rather than quietly papering over it.

"If the challenge is interesting and the impact is real, I'm all in."

Content Manager