Vishakha Fulare

Software Quality Engineer at Amazon, working on Alexa Smart Home Routines across Android, iOS, and FireOS. I build Python automation systems and the frameworks they run on.

Most of my recent work has been on a self-healing locator recovery framework I designed and shipped end to end — a modular system that detects when UI elements break at runtime and recovers automatically, with no human intervention. It's built around a plugin architecture with abstract strategy interfaces, so new recovery approaches (element-attribute matching, coordinate-based fallback, live page analysis through an LLM API) drop in without touching core logic. I also wrote the pytest plugin that makes it opt-in inside CI/CD pipelines. It runs on three platforms today and cut locator-related failures by roughly 90%, which removed a recurring manual fix cycle that was blocking releases.

Before Routines, I worked on Alexa Device OS and Bluetooth connectivity, where I designed automation architecture for multi-device execution spanning 3 operating systems and 35+ device types, and automated end-to-end coverage for the Alexa voice interface. Across both teams I've shipped 22 code reviews and around 8,500 lines of production Python.

On the research side, I published a paper in IEEE on TITAN, a dual-stream attention network for camouflaged object detection that fuses RGB and thermal modalities. It reached 64.85% IoU, a 120% improvement over RGB-only baselines, and I built a synthetic thermal data generation pipeline so the whole thing could be trained without thermal hardware.

I graduated with a B.Tech in Computer Science (AI/ML) from YCCE Nagpur with an 8.8 CGPA and Honours in AWS + DevOps. I care a lot about data structures and algorithms, clean design, and owning a problem from architecture through production. I'm currently looking for a Software Development Engineer role.