Works — Case Studies
Four real projects across four industries — a KGI monitoring dashboard, a data SaaS for 3,200+ machines, an enterprise migration off Salesforce, and a 14,284-member executive SNS built in Flutter. The honest account.
30min → 0
daily manual data entry
fully automated
← Pachislo SaaS
200+
migration bugs resolved
one by one
← Pachislo SaaS
32pd
to build the KGI dashboard,
after scope was clarified
← Telemarketing SaaS
Every project had a moment where the spec wasn't enough.
Undocumented workflows running in production. A mid-project handover nobody planned for. A four-company chain where every decision had to clear three approval layers.
That's when the real work started. This is that work.
Case Studies
10 project pages published in total. The four above are full case studies — development process, decisions, and outcomes. The rest are project summaries.
Telemarketing SaaS
0.19s
Over the SLA threshold.
The gap between "running" and "usable."
The numbers existed. Nobody could see them.
32 person-days. Requirements and development running in parallel across a four-company chain. SQL response hit 2.69 seconds — above the 2.5s SLA. Index restructuring, query splitting, cache redesign. That last 0.19 seconds was the actual project.
"No cross-tab analysis — just deliver these KGIs for this quarter." Scoping down once kept the team's priorities clear every sprint after that.
Data Engineering SaaS
3,200+
machines auto-synced daily.
Was 3 hours of manual work.
Every morning, three hours of someone's day disappeared.
Day one code review made it clear: this wasn't built for production scale. The previous vendor had left mid-project. We weren't inheriting a codebase — we were starting over.
Four months. Separate feature sets designed for three distinct user types: hall operators, data creators, and end users. 200+ bugs surfaced during migration. We resolved each one. "Cutting corners now becomes maintenance cost in three years" — that was the call.
Pharma & Healthcare
3mo
Full rebuild. 5 years of patient
data migrated. Shipped.
The workflow that ran the operation wasn't in the spec.
Three weeks into requirements, a fax confirmation flow came up for the first time. Not a single line in the official spec — but it ran every day in production. "Everyone knows it, so nobody wrote it down." That sentence kept coming back.
Full rebuild in React.js + Nest.js. Five years of patient records off Salesforce. Four functions built in-house: fax send/receive, CSV automation, document output, and physician sync. Four months, live February 2026.
Business SNS / B2B AdTech
14,284+
executives active in 30 days.
Japan's largest executive network.
Executives were connecting. The platform that sustained them wasn't.
BizOn! has the full feature surface of a modern social network — feed, DMs, matching, events, push notifications — built in Flutter. One Dart codebase, three platforms, 14,284 active executives in the first 30 days. Japan's largest executive networking community.
Adding an ad placement revenue layer triggered three simultaneous constraints: Apple's Steering regulation (pricing data in a Dart bundle becomes part of the iOS binary), GMO's redirect-based 3DS authentication, and Japan's Special Commercial Transactions Law. All three, on a live platform, coordinating iOS App Store, Google Play, and web in a single release window.
From the field — the LINNOEDGE team
Some projects we almost turned down. Too many stakeholders. Requirements still being defined mid-sprint. A previous vendor that left halfway through.
If you're hesitating to reach out because the situation feels too complicated — that's actually a reason to call, not to wait.
If that sounds familiar, let's talk.
— The LINNOEDGE team
We've delivered in pharma, regulated gaming data, and B2B SaaS — but the industry isn't the common thread. It's projects where undocumented workflows are running in production and the spec alone isn't enough. Our team is based in Vietnam with Japan-facing project management, which means we bring both cost efficiency and operational familiarity with Japanese business processes.
The three cases on this page each ran 3–4 months from kickoff to delivery. When requirements aren't finalized, we start with scope clarification rather than setting a deadline first. "When to start" matters less than "what we're actually building" — and getting that right early saves time later.
Yes. The regulated gaming data platform was exactly that situation — the previous vendor left mid-project and we inherited an unfinished codebase. We run a structured code review on day one and give an honest assessment: rebuild or continue from here. Complex handovers are where we tend to be most useful.
Contact
30 minutes is usually enough to figure out what's actually needed.