Origin Note
The job title changes. The instinct keeps showing up.
The thread through my work is pretty consistent, and I like walking into complicated systems, finding the shape underneath the mess, and making the next person feel less friction than I did.
That has taken a few forms, including migration tooling at HubSpot, security workflows at Tesla, design systems at Walmart, CPQ flows at Salesforce, and earlier research in computer vision. Different domains, same instinct.
I care about software that feels thoughtful in the small moments, where clearer states, fewer confusing paths, better defaults, sharper feedback, and codebases give teams room to move.
Make the system legible
Whether it is a UI migration or a product workflow, I try to expose the real structure so teams can make decisions without guesswork.
Keep the user in the room
I am happiest when engineering details disappear into an interface that feels obvious, fast, and hard to misuse.
Leave better trails
Docs, codemods, component stories, lint rules, and dashboards are all part of the same habit, which is to help the next person move faster.
Every role changed the domain. The underlying instinct stayed the same.
Research Wall
Before product systems, I was studying what machines can notice.
My early work sat at the intersection of computer vision, road safety, and autonomous vehicle support systems. That research background still shapes how I think about signals, edge cases, and feedback loops in product interfaces.
10th International Conference on Computing, Communication and Networking TechnologiesICCCNT 2019, pp. 1-6
A road-accident detection framework for traffic surveillance footage using Mask R-CNN object detection, centroid-based tracking, and trajectory anomaly analysis.
2020 IEEE 17th India Council International ConferenceINDICON 2020, pp. 1-6
A dashboard-camera support system for self-driving cars that tracks vehicles and analyzes speed, acceleration, and trajectory signals to detect accidents.
International Conference on Computer Vision and Image ProcessingCVIP 2020, pp. 237-248
A driverless-vehicle support framework that detects animals in dashcam video, checks lane obstruction, and tracks animal movement to reduce collision risk.
International Journal of Computer Information Systems and Industrial Management ApplicationsIJCISIM, Volume 13, pp. 182-191
A traffic signal and countdown timer detection framework for autonomous vehicles, combining Mask R-CNN object detection with image processing and RetinaNet-based timer recognition.
