Client · Delivered
3,905
Archive automation, LPG distributorship
Fourteen hours of manual triage to under one minute. 3,905 records reconciled, 116 fuzzy merges, zero rework.
- Python
- pandas
- Google Drive API
Read ↗
Kinyoubi Atelier & Co.
We build the systems that everything else depends on.
Engine, platform, API, and surface, shaped as one system. Most of it is invisible. All of it has to hold.
01 · What we build
Services for clients. Products of our own.
The engagement can begin at the engine, the platform, the interface, or the operating document. The system still has to hold as one.
- 01
Software Development
Full stack web and platform builds, across any modern language and framework.
- 02
Product & Platform Engineering
End to end product builds, shipped as one coherent system.
- 03
Technical Consulting
Architecture decisions, threat models, code reviews, audit prep, refactors, codebase rescues, migrations.
- 04
AI Powered Workflow Development
Orchestrators, idempotent pipelines, automation that survives production.
- 05
Enterprise & Institutional Systems
Custom ERPs, accounting modules, DBMS work, government and institutional websites.
- 06
Design and Brand Surfaces
Design systems, brand identity, visual languages, UI architecture for software products.
- 07
Documentation and Technical Writing
Runbooks, ADRs, security policies, audit ready documents, internal handbooks.
- →
Start a conversation
Share the constraints and the decision that cannot fail. We reply within one business day.
02 · What we build with
The working stack, on the desk.
Two honest sets. The build stack ships in production and traces to a published case study. The project tooling runs the practice behind the work and is not part of what we ship. Drag a block to move it; the desk is a live model, and the named links below carry every destination.
The build stack, in production
Project tooling, what we run the work with
03 · Selected work
Work that can be inspected.
One delivered automation, one regulated platform in progress, and one in house research system. The figures below belong to the delivered archive engagement.
- Delivered
- Archive automation for an LPG distributorship. Handed off, zero rework.
- In progress
- A regulated BFSI MIS platform. Foundation complete, phase two active.
- In-house research
- An in-house timetable engine, built end to end as proof.
Client · In progress
BFSI MIS platform
RBI aligned and DPDP built in. Foundation phase complete. Phase 2 is active. Client redacted under NDA.
- Postgres + RLS
- AWS Cognito
- AWS Bedrock
Read ↗
In house R&D
An in house timetable engine
A constraint solver, a multi tenant platform, a role aware front end, and a mobile scaffold, shipped end to end as proof.
- TypeScript
- Postgres + RLS
- Fastify
- Capacitor
Read ↗
04 · Founder note
We start with the decision the rest of the system depends on, and we record the constraints before any code. The path stays inspectable: policies, logs, tests, and runbooks sit beside the implementation, not in a folder no one opens.
What we hand over has to be something your team can run, audit, and extend without us in the room. That is the whole job.
05 · From the notes
How we think, in writing.
Working notes on the problems behind the work. Readable with no JavaScript and no web fonts.
- RLSData isolation
Multi tenant Postgres with row level security
One SQL clause and three operational disciplines. Skip a discipline and the clause stops protecting you.
read - SOLVERSConstraint solving
OR Tools, FET, and commercial timetabling
Three legitimate tools for one problem. Picking the wrong one is the kind of decision that compounds for years.
read - DPDPCompliance
DPDP for SaaS builders
The architect's reading of the Act, written for the schema design step where the obligations have to land.
read
Start with the system that has to hold.
Share the constraints, the current state, and the decision that cannot fail.