Joey Chakraborty

Unifying separate applications into one platform

Context
Araceli Biosciences’ imaging and analysis software was split across separate applications that users had to learn one at a time.
Role
I proposed merging four applications into one platform, won over engineering, product, and application science through three workshops, then with that team brought along the executive team, board, sales, and field.
Impact
We released EndeavorOS in September 2026, and the field team, the group most worried about retraining, was happy with it.
Built with
  • Figma
  • React
  • Claude Code
Timeline
15 months
Ship date
September 2026
EndeavorOS home screen, recreated with sample data.

The situation

The pain points of a disjointed suite

Araceli’s software was four separate applications. Each had grown independently, so terms, interactions such as how wells are selected, and the look of buttons, colors, and icons all differed from one application to the next.

Our service team had to take users through a separate training session for each application, which made training very long. People also alt-tabbed between them as they moved through various tasks. For a company that sells usability, that wasn’t acceptable.

Concept view. Four applications, unified with shared principles, structure, terms and design, into a single application.

The work

The case for one platform

In early 2025, I proposed merging the four applications into a single application: one platform with one information architecture, one design language, and workflows that cross what had been separate products. The first response was lukewarm, because people couldn’t imagine beyond what we had and how customers already worked with it.

It had two costs: some features would pause while we rebuilt, and everyone who knew the old software would have to relearn it. I argued that one consistent application would be more intuitive, so retraining would happen once and training would get shorter after it.

A workshop board of six design principles on sticky notes, with votes, ranked 1 to 3.
Workshop exercise: the team voted on the principles that should guide the platform’s design. Trust and predictability ranked first.

The work

A different yes for each group

I pulled our software engineering lead aside first to talk through what unification meant for building and maintaining it, and he was on board quickly. I then ran three workshops with the product managers, the engineering lead, and our application scientist, walking them through old and new workflows side by side, and voting together on the principles that should guide the design.

Product managers needed many mockups (I built them in Figma), proof that deployment would stay flexible, and time to decide the paused features were worth it.

With that team behind it, we took the idea to the executive team and the board. I presented it with our software product manager at the company sales kickoff. Field and sales were the most hesitant because of retraining, but came around once they saw the concepts.

The EndeavorOS catalog in its dark scheme: a list of plates with a Plate filters dialog open, filtering by date range, well format, channels, and analysis status.
Me at SLAS, standing beside an Endeavor Ultra instrument and a monitor showing an early concept of EndeavorOS.
Left: The catalog, one place to find and manage plates, protocols, and other named entities. Right: Me at SLAS 2026, showing an early concept of EndeavorOS.

The outcome

Shipped, with the field team on board

Building began after the workshops, and I moved from Figma into code with Claude Code. I design in my own branch, which our software engineers pull and connect to the back end. EndeavorOS shipped in September 2026.

The field team was happy with what we shipped, including the new catalog screen I rallied for, a single place to manage named entities like plates and protocols. The platform now has room for new applications and add-ons without splitting apart again, and early customer feedback is coming in.