Welcome to my little corner of the web, where I share my musings, interests, and biases.

I’m a full-stack data scientist with a Doctorate in Astrophysics. Originally from the Himalayas, I now call Australia home. I’m the product of a humble upbringing and a lifelong journey of learning, unlearning, and growth. My path has been shaped by the unwavering support of my family, the generosity of brilliant mentors who continue to inspire me, and with a great deal of hard work.

Over the years, I have built deep expertise in Marketing and Customer Data Science, delivering suites of predictive and prescriptive machine/deep learning and statistical models (both Freq and Bayesian paradigms) — ranging from churn and purchase propensity, attribution models to generosity sensitivity scoring, and even automated case-note triaging. I have designed, delivered, and deployed company-wide experimentation platforms for VGW and Sportsbet, probably my academic background in discovery have positioned me for this. Most recently, I’ve been providing thought leadership in state-of-the-art Reinforcement Learning, Causal Inference, and pioneering MLOps capabilities, including Feature Stores.

Earlier in my Post-doctoral Position at UWA, I developed a physically motivated, unsupervised clustering algorithm in Python, trained on data from the GAMA Galaxy Redshift Survey [1]. This algorithm refined galaxy distance (redshift) measurements by associating them with their corresponding galaxy clusters.

Similarly, during my doctorate at USyd I studied and used the kinematic properties of the Milky Way and estimated its total mass to be under a trillion solar masses!! The papers I co-authored with my mentors have since earned gratifying citations worldwide [2] and [3].

Out of academia, my interests grew on all three core operations of Data Science, namely,

• Machine Learning Ops: End-to-end ML Model Delivery from Scoping to Productionisation, Drifts, Decay and Monitoring

• Experimentation Ops: Stat and Commercial Significance, A/B Testing, Causal Inference

• Scientific Excellence Ops: Model Delivery Life Cycle, Code/Peer Review all within CRISP-DM format.

I code in python mainly, love to contribute to open source Data Sciency initiatives, and also support business’s internal shared code-base.

All words expressed in this blog, including diary entries, are my own and reflect my personal views and biases—they do not represent the opinions of my current or past employers. This blog is licensed under the MIT License.