
Ro-ee Tal
Hi, I'm Ro-ee, welcome to my blog!
I am currently an entrepreneur and AI researcher based out of Tel Aviv. I decided to leave New York and move to Tel Aviv to be closer to my family, and to avoid breaching the terms of my L1 visa status in the US to work with some friends on a fintech startup. Our vision was to provide alternative credit-based solutions for tenants and landlords to make moving in and out of apartments drastically easier for all parties.
Over the 7 months we worked on it, we were able to secure a letter of intent from a regulated bank to issue the financial product and onboard several landlords representing approximately 14,000 rental units to pilot the solution. As we worked through the bank's proposed terms, we concluded that retaining the credit risk left too little margin for error, while a capped pilot and long lease commitments would make the model slow to validate and difficult to adjust. We chose to wind down before raising external capital or signing definitive program agreements.
Before founding Parent, I was an applied scientist at AWS based in NYC, where I focused on creating language models and agentic systems that perform robustly for business applications. Much of my research focused on desinging and improving RAG systems, evaluating the quality of conversational AI, and building tools to observe the behavior of agentic systems at scale. Some of my work contributed to the early development of the StrandsAgent harness, the Bedrock AgentCore evaluation and observability managed services, and the Q chatbot in the AWS console. My earlier years at Amazon were spent on improving technical customer support: from domain-specific translation for cross-lingual support to ranking and recommending millions of Trusted Advisor risks across entireprise organizations systems.
Before joining Amazon, I studied engineering physics at the University of British Columbia. There, I built robots, simulated physical systems, worked in a quantum devices lab, designed software, and did some computer vision research for health-care and the forestry industry. Towards the end of my studies, I spent a semester at ETH Zurich where I focused on deep learning. I was born and raised in Johannesburg, South Africa, and, from what I can remember, spent all my time there in the sun before moving to Canada.
I decided to create this blog for a couple of reasons, the main one being to share some of my thoughts, ideas, or projects with people who may find what I have to say interesting or would be willing to provide some constructive feedback. But I also hope this blog will serve as a platform to inspire personal growth. I want to get better at writing more clearly, more effectively, more efficiently... and to become comfortable expressing myself imperfectly. After all, learning is a continuous and holistic journey; and doing it quickly requires this sort of temporal vulnerability. And maybe, over time, it will paint an interesting picture of my journey.