Craft2 min read
Learning AI, on purpose
AI is moving faster than any of us can passively absorb. So I stopped trying to absorb it and started learning it deliberately — as a backend engineer, on purpose, a piece at a time.
Backend Engineer · Microservices · Distributed Systems
Engineering scalable systems. Grounded in the peaks.
I design and build the invisible, scalable backbone of modern applications — microservices, enterprise workflow automation, and IoT infrastructure. At Racanaa Energy I own backend systems that quietly manage 25,000+ IoT assets across 500+ enterprise sites. Lately I've been deep in building data, video-streaming, and image-processing pipelines with a sprawling toolbox of technologies — and teaching myself AI, on purpose. When I'm not optimizing database schemas or deploying to AWS, you'll find me setting up plays on the volleyball court, exploring the peaks of Uttarakhand, or capturing the world through a camera lens.
Latest writing
All posts →Engineering2 min read
Pipelines all the way down: data, video, and pixels
My favourite kind of engineering is plumbing. Give me a firehose of data, a stream of video, or a pile of images and a pipeline to push them through, and I'm happy. Here's why — and the sprawl of tools it takes.
Engineering2 min read
Designing a backend that manages 25,000 IoT assets
REMA configures real hardware across 500+ enterprise sites — meters, sensors, controllers, gateways. When a bad write can mis-configure a physical device, the backend's job stops being CRUD and starts being safety.
Life & Peaks1 min read
Keeping the ball alive since school
Volleyball is my ultimate reset button. It also, surprisingly, taught me how to debug a production crash at 11pm without panicking.
Craft1 min read
Automating approvals without losing the humans
A workflow engine for enterprise service requests — multi-level approvals, SLA tracking, SOP scheduling across time zones. The hard part isn't the automation. It's deciding what stays human.
Get new writing in your inbox
No schedule, no spam — just new essays when they’re ready. The slow web, delivered slowly.