Women in Orange BG

Stellaris Platform

AI Intelligence / Internal Build

© 2026

Women in Orange BG

Stellaris Platform

AI Intelligence / Internal Build

© 2026

Women in Orange BG

Stellaris Platform

AI Intelligence / Internal Build

(01)

(Read More)

© 2025

An AI talent management platform, built in-house to augment human hiring ops with automated pool management and coordination

Men Red BG
DAYDA built an internal platform to hold everything the studio knows about the people it hires. Applications arrive from five different places, get read and structured on the way in, and land in one pipeline that a Claude-powered Slack agent and other software can both work from.Stellaris began as a candidate list and became the studio's talent record. Applications come in through the platform, public job forms, webhooks, an AI assist field and Slack. Whichever route they take, they land in the same pipeline under the same rules, with resumes read and structured on arrival.

(Five ways in, one record)

Candidates reach us through the platform itself, an AI assist field, public application forms, webhooks from other tools, and the Slack agent. Each route had to write to the same pipeline with the same validation, so that no single route becomes a private version of the truth. We built one ingest path underneath all five surfaces and put every write through it. The kanban board is the record itself, so there is nothing to reconcile afterward.

(Unstructured applications, structured pipeline)

Applications turn up as PDF resumes, portfolio links and LinkedIn profiles. Typing them into a spreadsheet was the standing cost of every open call. Claude now reads each submission as it arrives, pulls out the fields the pipeline needs, and stages the candidate for review. A person still confirms the result before it moves, which keeps the parsing accountable.

Applications turn up as PDF resumes, portfolio links and LinkedIn profiles. Typing them into a spreadsheet was the standing cost of every open call. Claude now reads each submission as it arrives, pulls out the fields the pipeline needs, and stages the candidate for review. A person still confirms the result before it moves, which keeps the parsing accountable.

Men Red BG
stellaris

Public applications reach a reviewable candidate card without anyone typing them up, and that is where most of the reclaimed time sits. The Slack agent answers questions like who we know with Houdini and water simulation experience, reading and writing the same database as everyone else. Its edits appear in the activity log the way a person's do, so the record stays auditable regardless of who or what touched it.The build also gave us a methodology that we reuse. One gated platform and one public domain served by the same middleware. One scoped-key primitive for every external caller, so adding an agent is a permission decision rather than a new pathway. An MCP server so AI tools read the system through those same permissions. We now deploy that pattern for teams whose working knowledge is sitting in inboxes, DMs and spreadsheets.

58%

Of intake arrives fully parsed and staged with no manual work

Redefining minimalism through material authenticity

and design order. Forma moves beyond simple.

5 > 1

DATA STREAMS FEEDING ONE PIPELINE UNDER ONE SET OF RULES

Redefining minimalism through material authenticity

and design order. Forma moves beyond simple.

Men Red BG
Men Red BG
Men Red BG
Men Red BG

DAYDA.STUDIO

(02)

(Portfolio)

© 2025

(02)

(Portfolio)

(03)

(Frequently Asked Questions)

WHO IS ACTUALLY DOING THE WORK; Human or AI?

WHO OWNS THE WORK, AND HOW DO YOU HANDLE IP?

CAN YOU WORK INSIDE OUR TEAM AND ON OUR existing STACK?

WHAT HAPPENS TO CUSTOM workflows and TOOLING WHEN THE PROJECT ENDS?

DO WE HAVE TO COMMIT TO A FULL PROGRAMME?

(03)

(Frequently Asked Questions)

WHO IS ACTUALLY DOING THE WORK; Human or AI?

WHO OWNS THE WORK, AND HOW DO YOU HANDLE IP?

CAN YOU WORK INSIDE OUR TEAM AND ON OUR existing STACK?

WHAT HAPPENS TO CUSTOM workflows and TOOLING WHEN THE PROJECT ENDS?

DO WE HAVE TO COMMIT TO A FULL PROGRAMME?

(03)

(Frequently Asked Questions)

WHO IS ACTUALLY DOING THE WORK; Human or AI?

WHO OWNS THE WORK, AND HOW DO YOU HANDLE IP?

CAN YOU WORK INSIDE OUR TEAM AND ON OUR existing STACK?

WHAT HAPPENS TO CUSTOM workflows and TOOLING WHEN THE PROJECT ENDS?

DO WE HAVE TO COMMIT TO A FULL PROGRAMME?