​Location: visibile onxe registration approved (VC Hub in Toranomon, Floor 5!).

​Summary

​Join us for an evening with Tractable, a leading AI-first InsurTech company, as they share insights on applying machine learning and computer vision in the insurance industry.

​CEO Venkat Sathyamurthy will discuss Tractable’s approach to entering the Japanese market and building trust through practical, high-impact AI solutions. He'll be joined by senior AI researchers who will detail how Tractable develops and deploys scalable ML systems for auto and property insurance.

​​Schedule

​18:00 Doors open

​18:30 - 18:50 Succeeding in Japan with an AI startup (Venkat Sathyamurthy, CEO Tractable)

​18:50 - 19:10 AI Agent for Insurance - Changing Industry Together (Toru Tanaka, CEO Milize)

​19:10 - 19:35 From Lab to Live: AI system evaluation in practice (Filippos Siozopoulos)

​19:35 - 20:00 Rapid Property Damage Analysis with Applied AI (Elham Asadi)

​20:00 - 21:00 Networking

​21:00 Event ends

​Talks

​Talk 1: Succeeding in Japan with an AI startup

​Speaker: Venkat Satyamurthy, CEO Tractable

​Abstract: He will unpack how his computer-vision AI company, renowned for real-time vehicle damage assessments, won over Japan’s cautious insurance and automotive sectors. Drawing on 25+ years at Microsoft, Adobe, and beyond, Venkat will reveal how Tractable built trust through precision, reliability, and seamless integration with local workflows.

​He’ll explain why Japan’s unique mix of technical sophistication, aging demographics, and enterprise openness makes it a global hot spot for AI innovation, and share practical strategies for localizing product-market fit, navigating enterprise sales, and forging lasting partnerships. Finally, Venkat will offer his perspective on AI’s next frontier and why Japan will be central to that evolution. Essential insights for founders, investors, and technologists aiming to thrive in Japan’s AI landscape.

​Bio: Venkat brings over 25 years of leadership in enterprise tech with roles at Microsoft, Adobe, and DocuSign. A firm believer in “outcome, outcome, outcome” philosophy, driving the deployment of Tractable’s AI-powered computer-vision solutions across auto-insurance claims, vehicle recycling, fleet management, and damage assessments to deliver measurable business impact. Known for his customer-centric approach - “meeting customers where they are” - he focuses on solving real-world challenges today rather than chasing futuristic buzzwords.

​Talk 2: AI Agent for Insurance - Changing Industry Together

​Speaker: Toru Tanaka, CEO Milize

​Abstract: We will introduce how Milize is using LLM–driven agents to transform insurance and financial services. Leveraging partnerships with leading AI platforms such as SmythOS, Dify, and Ryoma.ai, Milize has developed multi-agent systems capable of autonomously handling complex workflows, from financial planning and claims analysis to compliance checks and customer engagement.

​In this session, we will outline the design of Milize’s AI Agent platform, where specialized agents collaborate to understand problems, analyze data, generate solutions, and deliver advice with high precision. We will share real-world use cases in insurance and demonstrate how agentic workflows can enhance operational efficiency, improve customer experience, and enable rapid service innovation. Attendees will gain practical insight into building scalable, domain-specific AI agents that meet the demands of regulated industries.

​Bio: Toru Tanaka is a seasoned entrepreneur and business leader with a strong track record of founding and scaling companies across diverse industries, including financial technology and food services. As the CEO of MILIZE, Inc., he leads the development of cutting-edge AI and fintech solutions that support individuals and financial institutions in achieving smarter financial decisions.

​Talk 3: From Lab to Live: AI System Evaluation in Practice

​Speaker: Filippos Siozopoulos (Fil), Senior Applied Researcher

​Abstract: Ready to retire your sandbox models? Join Fil in this deep dive into evaluating AI systems and learn how to transition AI prototypes into robust, production-ready systems.

​You’ll learn how to spot and squash “Trust Eroding Errors,” collect and curate data efficiently, and make clear go/no-go decisions. He will cover practical testing strategies, from sampling live claims data to setting up automated health checks, and show how tools like Weights & Biases streamline monitoring and feedback loops.

​You’ll walk away with a playbook for turning prototypes into reliable, scalable AI solutions that insurers and customers can trust. Perfect for anyone into taking ML models from the whiteboard into real-world workflows.

​Bio: Filippos “Fil” Siozopoulos is a Senior Applied Researcher at Tractable, where he leverages expertise in NLP, deep learning, and computer vision to optimize insurance claims and damage-assessment workflows. Prior to Tractable, he advanced cutting-edge modeling as a Senior Machine Learning Engineer at Samsung Research UK. Fil holds an MSc in Machine Learning from UCL and a master’s in Electrical & Computer Engineering from the National Technical University of Athens. He excels at translating research into resilient, high-impact AI solutions for the insurance industry.

​Talk 4: Rapid Property Damage Analysis with Applied AI

​Speaker: Elham Asadi (Ellie), Senior AI Applied Scientist

​Abstract: When every minute counts after an accident or natural disaster, AI can turn a simple smartphone photo into a reliable damage appraisal, fast. This session dives into the core components of a scalable AI pipeline for property damage assessment: image-quality verification, object and damage detection, and cost-estimation modeling.

​You’ll learn about best practices for curating high-value datasets, handling edge-case errors, and embedding real-time evaluation to maintain accuracy across varied conditions. Gain practical insights to build scalable, end-to-end AI systems for insurance, disaster recovery, and field-service applications.

​Bio: Elham “Ellie” Asadi is a Senior Applied Scientist on Tractable’s Property team, where she’s spent over three years developing computer-vision AI for property damage assessment in insurance. She brings 15+ years of machine learning and computer-vision expertise across aerospace and automotive, including AI for autonomous vehicles at HORIBA MIRA. Elham holds a PhD in Aerospace Engineering from the University of Bristol and completed a postdoctoral fellowship at the University of Cambridge.

​Organizers

​Ilya Kulyatin: Fintech and AI entrepreneur with work and academic experience in the US, Netherlands, Singapore, UK, and Japan, with an MSc in Machine Learning from UCL.

​Atul Anand: Tech evangelist and AWS Community Builder, currently building AI products in the Insurance domain, in Tokyo, Japan.

​Supporters

​Thanks to GHOVC and Tractable for their support of this session!

​​Our Community

​Tokyo AI (​​TAI) is the biggest AI community in Japan, with 2,500+ members mainly based in Tokyo (engineers, researchers, investors, product managers, and corporate innovation managers).