Product Growth
Lifecycle and conversion work that connects user needs, pricing, onboarding, and post-launch measurement.
Product-minded Fintech Engineer
I build and operate fintech systems at the intersection of payments, risk/compliance, product growth, and operations tooling. My work turns ambiguous customer and internal workflows into reliable, monitored systems that keep improving after launch.

The common thread is owning ambiguous fintech problems end-to-end: define the customer or operations pain, ship the system, monitor the rollout, and keep improving the numbers after launch.
Lifecycle and conversion work that connects user needs, pricing, onboarding, and post-launch measurement.
Regulated workflows, monitoring, review tools, and operational controls for sensitive fintech surfaces.
Payment and transaction experiences designed with backend reliability, observability, and supportability in mind.
Internal workflows, support tooling, dashboards, and automations that make teams faster and safer.
Sensitive implementation details and private metrics are summarized at a high level here; the focus is scope, ownership, and engineering judgment.
2025 - Present
Lead engineer on the Infrastructure team, serving as directly responsible individual across parallel workstreams spanning payments, customer-support architecture, and internal platform and API engineering. Designed and built Atlas HQ — a Next.js internal operations platform replacing Retool — including authentication, role-based access control, and migration of all user, risk, dispute, and activity tooling, with rollout to 100% of the support and operations org. Architected the company's end-to-end customer-support system on Intercom with team-based routing, retention and escalation logic, cancel-flow integration, and an LLM library for understanding customer conversations. Owned integration of new payment rails — debit-card instant deposits and a cash deposit network — covering ACH file generation, reconciliation, chargebacks, 3DS, fraud controls, and partial-authorization deposits.
2023 - 2025
Built and shipped core consumer-facing features in Flutter and TypeScript on Firebase — new-user onboarding, retention, and Plaid bank-connection flows — later transitioning ownership to dedicated teams after moving to Infrastructure. Drove growth of the paid Annual Plan subscription from initial launch into a core revenue stream through pricing UI experiments, new upsell entry points, and funnel analysis. Led the Spanish-language initiative end-to-end — localization quality, KYC decline-rate reduction, ITIN support, and Spanish-market campaigns — improving conversion and approval rates for Spanish-speaking customers. Delivered KYC and OFAC compliance infrastructure including OFAC monitoring, a risk-engine migration, eCBSV/CIP integration, and dispute and Metro 2 tooling, alongside reliability work including an activity-tracking framework, monitoring dashboards, and on-call processes.
Jun 2020 - Dec 2023
Built and operated large-scale payment infrastructure handling millions of requests daily using Java and the Spring framework. Designed and implemented payment failure-recovery components spanning HTTP clients, ActiveMQ messaging, batch jobs, caching, and Oracle databases, with Vue-based internal UIs.
Apr 2019 - May 2020
Built a full-featured web client for a tutor scheduling system that allowed students to book sessions and administrators to audit schedules and activity. Implemented the UI with React and JavaScript and integrated it with a PHP backend.
May 2019 - Aug 2019
Built a React, Redux, and TypeScript web client for analyzing online advertising performance. Also supported data science research aimed at improving ad selection model accuracy.
Jan 2019 - May 2019
Applied YOLOv3 to TEM alloy video frames to detect, count, and track material defects. Built an analysis workflow that measured defect generation and annealing, achieving 0.95 F1 with low-latency detection.
Westcliff University
University of Wisconsin-Madison
University of Wisconsin-Superior
Waseda high school (Japan)
Organization-published engineering articles on fintech systems, internal platforms, and product reliability.
Why modeling a financial product's status as explicit, explainable states improves reliability, support, and user trust.
How short-lived, support-scoped session tokens help WebView support surfaces stay fast, secure, and recoverable without exposing a mobile app's primary credential.
Explains how fintech internal tools should be treated as risk systems, with action-level permissions, approval workflows, audit trails, request IDs, and idempotent execution for sensitive operations.
Notes on engineering, lessons learned, and ideas explored in public.
The shogi AI on my blog was weak. Weak enough that I — an amateur 2 dan — beat it comfortably, and weak enough that it lost every single game to climbing silver

Inside Personal Memory — MCP, OAuth, PostgreSQL, and pgvector for Cross AI Long Term Memory (Part 2) Part 1: I stopped re explaining myself to every AI explai

Every new chat started the same way: I would explain my job, my goals, what had been bothering me lately, and everything I had already tried. I built Personal M

Production-ready utilities built to solve real-world day-to-day problems.
Smart group payment settlement. Automatically finds the optimal transfer plan for complex bill splits.
Smart shopping lists with budgets, categories, sharing, and recurring item reminders.
Multiplayer quiz platform with 400+ questions, daily challenges, custom quizzes, and live scoreboards.
Drag & drop or paste markdown files for instant full-screen preview with GitHub Flavored Markdown support.
Plan fictional railways from national trunk lines down to zoomed station layouts for all 47 prefectures.
Extract tasks from screenshots or text with AI. Manage your todo list with smart input.
麻雀の素点、ゴルフ、ボドゲ会など — 日付ごとの最終点を記録して累計を追跡。コードで友達と共有できます。
More practical tools are on the way.
Personal, community, and engineering builds with context on the problem, technical choices, and outcome.
Founder & Organizer — JTPA-supported Bay Area AI Study Group
Beyond my fintech work, I founded and lead the Bay Area AI Study Group, a JTPA-supported AI community and self-built platform where Bay Area engineers share projects, articles, and lessons learned.
Plan and run the monthly meetup end-to-end: agenda, speakers, venue, and check-in.
Designed and built bayarea-ai.com from scratch with Next.js and Firebase.
Review and showcase member projects, articles, and tips submitted to the community hub.
Lead a Japanese-speaking engineer community around AI in partnership with JTPA.
The community runs in Japanese for Bay Area engineers; the platform and event information are also available in English.
Articles and interactive CS labs for learning by reading, running, and checking understanding.
Playable experiments focused on UX, performance, and game logic design.

Senior Fintech Engineer
Hi, I'm Yudai — a senior software engineer based in San Francisco with 6+ years building production fintech systems across payments, risk, compliance, growth, and operations. At Atlas and PayPal, I have owned payment integrations, KYC/OFAC and dispute workflows, support and risk tooling, monitoring, and customer-facing activation and retention experiences. I am strongest when the problem is cross-functional: turn unclear business goals into a shippable system, coordinate rollout details, instrument the results, and keep improving the customer and operations experience. Outside of work, I lead an AI meetup community for engineers and keep exploring how machine learning fits into everyday products. I'm pursuing an MS in Computer Science at Westcliff University while looking for the next role that blends deep engineering with product ownership.