About Me
I appreciate your interest in learning more about me. While this space offers a glimpse, I believe in the value of genuine professional connections. I welcome you to reach out via my social channels, where we can engage in meaningful discussions about our experiences, insights, and potentially explore collaborative opportunities in our field.
I currently work as a GPU Developer Advocate at NVIDIA with the OpenHackathons Group, where I help developers, researchers, and industry teams move AI and scientific workloads closer to production-grade GPU acceleration. A large part of my work involves mentoring hackathon teams, building technical enablement material, and helping people reason through performance, deployment, and architecture choices.
Across 7+ years in data science, machine learning, and developer advocacy, the core thing I have learnt is to solve real problems with systems that can survive contact with messy data, changing requirements, and real users. I like building fast feedback loops: get the first useful version running, measure what matters, then keep tightening the solution.
My current technical focus is Generative AI beyond the demo layer: RAG, LLM inference, model deployment, multi-agent workflows, and agentic AI applications that can use tools, retrieve knowledge, coordinate tasks, and fit into practical business or research workflows. I also continue to work across large-scale data processing, traditional machine learning, deep learning, and GPU-accelerated Python workflows.
Before NVIDIA, I was a Manager & Lead Data Scientist at Shiprocket, where I built data-heavy and scalable ML pipelines across Tabular ML, NLP, logistics intelligence, fraud detection, address intelligence, and time-series forecasting.
Before that, I had a short stint at Whitehat Jr., an EdTech company, where I worked on clustering, segmentation, data crunching, and dashboards used by CxO teams. I began and evolved into data at EXL, where I transformed slow, monolithic projects built in proprietary tools into faster, modular, interpretable, and user-friendly solutions.
Some of my key projects are listed here. I also undertake courses and certifications to keep myself up to date with this rapidly evolving domain.
I have no defined interests 🙈 and take up and try different things. Trying to keep a memoir here. Besides that, I am a watch aficionado and 🍺 buff.
I also try to maintain some lists for my:
On most platforms, you’ll see me going by the alias boi-doingthings.