Michael Senkow
I Build Machines
Focus was around SQL work but, tasks done here led into what became their current Meta AI image creation flow.
initial creation of daiquery notebooks, a collection of SQL cells for various reasons
The initial goals in this group were focused around the creation of a “SQL cell based notebook” as a way to better connect complex queries for storage and understanding by our Data Scientist / Data Eng / Software Eng users. The initial work was focused around discovering the everyday problems in this org. Connecting for feedback over the workplace groups, interviewing users around their issues, and generally understanding the product and similar ones in the field.
‘Cells’ as a concept became an oddly large focus for the next few years, with different types being added, visual treatment shifting to accomodate the varying additions, and some good discussions occuring with analogous notebooks outside the company.
discovery of common feature sets with nearby applications
Through expansion of this groups product though, we discovered commonalities with other product spaces within infra. We were an SQL notebook, but there was already a Python based notebook that had similar if slightly different features. Daiquery came with a file cell while bento came with a file panel.
This impacted and changed discussions around a wide range of areas of the product. Here for example the structure of the cells and the interactions that occurred around them, what functionality existed here versus existing in a panel region, became part of the discovery and goals process.
Example figma board with some of those studies. Feel free to explore!
A feature concept that didn’t get built, but was aimed at helping a common issue brought up in research around finding the various cell types amongst the complexity.
Some of the more finalized ‘cell’ forms, between SQL/Python/Data Exploration types.
expansion into additional application concepts
Early discussion were with cross functional PM / ENG / DS around regions that these two products were missing that they could benefit from, that they saw in other applilcations. Concepts like the an expandable plug-in/app store, education section, better sharing etc.
crafting and selling the org on a ‘universal shared notebook’
These evolved into final prototypes, connecting groups and planning for the next half.
Deck from that story-telling period.
company changes, shift of focus while continuing to support users
Some company upheavals occurred and I had a learning period around accepting certain losses in the larger goal, but keeping them on the back burner for later dates.
Our original story had the goal of expanding the notebook outside of the purely technical sphere and benefiting wider ranges of users but that was reduced without an education section or better documentation.
Focus resumed back to more purely technical with a few research discussions broaching into how PMs or slightly more technical designers used the product.
People were using it purely for notes and this added some life to a workflow around upgrading the text cells into more of a proper ‘doc-esque’ cell-type.
ai as a chatbot vs ai as a static cell type
AI started to become part of almost every discussion the last two years. How can we surface join recommendations or potential data-objects that relate to what a user is doing in either SQL or python? Do they want to see these recommendation in-situ, or via a helpful bot? Do they want to compare and contrast results or just have the code produce results and let them get about their day.