Taipy’s story is a beautiful one. The founders and creators of Taipy, Vincent Gosselin and Albert Antoine, come from extensive companies backgrounds developing advanced applications in AI and AI systems for big companies (ILOG and IBM)
These applications were developed in Java, and they wanted to move to Python, which is becoming the mainstream language.
However, creating ETL and pipelines in Python by datascientists using different libraries is quite vast and challenging to industrialize and not easy to use by the end user.
Indeed, according to Venturebeat, 87% of data science projects don’t make it into production because moving to full applications is time-consuming and too expensive.
State of play
For over ten years, data scientists have been spoiled with a flurry of great software to help build AI algorithms.
From the essentials :
To the AI libraries:
And the platforms:
Problem
Only 20% of pilots make it into production.
However, little is available to help data scientists bring the algorithms into end-users hands, which explains this high failure rate.
End-users needs
🙉 Business users will not just accept the output of an algorithm, however brilliant it can be.
👉 They will require a customized and highly interactive interface.
👉 They will want to perform what-if analysis and scenario management.
👉 They will need to monitor business KPIs.
👉 They will also want to collaborate with other users.
🚀 This is precisely what Taipy brings to the table with its two components: 𝗧𝗮𝗶𝗽𝘆 𝗖𝗼𝗿𝗲 and 𝗧𝗮𝗶𝗽𝘆 𝗚𝗨𝗜.
Why Taipy was born
👉 Develop using their favorite IDE,
👉 Their only web application builder fully compatible with Notebooks,
👉 Benefit from multi-user sessions,
👉 Program with code autocompletion,
👉 Bring graphical components from third-party libraries with the API extension,
👉 Deal with large data visualization.
👉 Build their pipelines graphically in no time with Taipy Studio,
👉 Bring pipelines’ orchestration to the next level with :
-
- Skippable tasks,
- Data nodes,
- Data scoping,
- And scenarios,
👉 Compare scenarios and track their performance over time without effort,
👉 Surprisingly, pipeline versioning didn’t exist so far! Well, it does now with Taipy…
Check out all the reasons Taipy was born and what features it comes with.