Hey! I'm Katya and I'm building AI flood detection models at Floodbase where we help communities prepare for and respond to floods. I love walking, reading and lifting (not at the same time).
Timeline
-
2026 – now
I'm a senior machine learning engineer at Floodbase.
I trained and shipped a model that detects flood extents globally over a 24-year history, and now I'm working on a specialized version of that model for Europe. We can map flooding under clouds by combining radar data and optical imagery with coarser, more frequent observations. These models are used for flood parametric insurance and event response worldwide.
-
2024 – 2026
As a senior software engineer, I shipped our US flood model. I designed and built validation workflows that proactively catch quality issues by tracking high-trust data sources and LLM-scraped signals. I also saved the company ~70% in recurring cloud costs.
-
2022 – 2024
As a software engineer, I worked on our data platform that helps clients respond to disasters and finance climate adaptation and recovery. I designed and deployed a high-throughput zonal statistics API used globally over a 42-year history.
-
2016 – 2021
I graduated with an Honours Bachelor of Computer Science (Co-op) from the University of Waterloo. I also completed a specialization in Human-Computer Interaction and a minor in Cognitive Science.
Co-op software engineering and research internships while at Waterloo
-
2021
On Datadog's cloud security team, I implemented live tracking for over 100,000 pieces of cloud infrastructure, as well as automated checks that keep every port in the company off the public internet.
-
2020
On Datadog's open source community team, I built a Kubernetes CTF (security challenge) for ~1700 people at AWS re:Invent.
-
2020
As an early-stage engineer at Abnormal AI, I built and launched one of their first APIs for handling social engineering email attacks. I also developed and launched a Kafka service that buffered throughout a multi-hour downstream outage, then cleared the 2.5-million message backlog in thirty minutes.
-
2019
I completed a research fellowship during my study abroad term at the Singapore University of Technology and Design. I developed an online framework for a hybrid timed-automaton + Bayesian network anomaly detection model, preventing cyberattacks against a water treatment plant by monitoring its sensors for anomalies.
-
2019
At Toast, I worked on an event-sourced service that synced data between point-of-sale devices in restaurants, so devices could drop off the network and rejoin without leaving the system in a bad state.
-
2018
At North, I prototyped new interactions for Focals smart glasses alongside HCI researchers. North was acquired by Google six months later.
For my full work history, visit my LinkedIn.
Talks
I designed and delivered three workshops on security engineering for Hack the North, Canada's largest hackathon.
-
Sep 2021
Deep Dive into Security Engineering
-
Sep 2021
Large-Scale Security Engineering
-
Feb 2021
Intro to Security Engineering