Computer engineer building probabilistic machine learning systems for aviation safety — currently collaborating with Embry-Riddle's College of Aviation on ADS-B surveillance deployment in Nepal, with CFIT detection research now under peer review at a Springer Nature journal.
I'm a computer engineering graduate from Pokhara University with a deep interest in the intersection of machine learning and aviation safety. My capstone research, "Extending GPWS Through Probabilistic Machine Learning for Early CFIT Detection in Challenging Approach Environments," made one thing clear — this is exactly the problem space I want to spend the next chapter of my career in.
Nepal's mountainous terrain makes its aviation environment one of the most demanding in the world. I believe data-driven approaches can meaningfully shift the odds, and I'm working toward a Master's in Applied Aviation Safety to build the technical foundation to do that properly.
Beyond research, I've spent time teaching computer science at the secondary level — an experience that sharpened how I explain complex ideas clearly, which I think makes for better research and better collaboration.
Since September 2025 I've been a volunteer research collaborator with Embry-Riddle Aeronautical University's College of Aviation, working on ADS-B/MLAT surveillance deployment across Nepal — from CAAN regulatory cross-referencing for receiver siting to standing up and validating a live surveillance feed on the ground.
"Every safe flight is a system working as designed. My purpose is to help build an aviation system where safety is the standard, not the exception."
From ML pipelines to classrooms — a well-rounded foundation.
Research and engineering at the intersection of ML and real-world safety.
Research into probabilistic ML methods to improve aviation safety systems, targeting Controlled Flight Into Terrain (CFIT) — one of aviation's most persistent fatal accident categories. Built a DEM-bounded simulation framework and applied rare-event modelling and temporal sequence modelling (TCN) to push the boundaries of ground proximity warning capability, evaluated against operational scenarios at Pokhara International Airport. Currently under peer review at a Springer Nature journal.
End-to-end ML pipeline processing 25,000+ race entries to predict Formula 1 championship outcomes — with rigorous feature engineering and interactive visualisation.
AI-driven system that classifies air traffic control voice commands in real time using DistilBERT and speech-to-text integration.
A research-focused toolkit built around Python, with depth in ML and data analysis. Dial it up like a CDU.
Building toward postgraduate study in aviation safety.
Research findings written up and put in front of peer review.