Hi everyone,
TLDR: Please let me know if you are interested in joining a webinar(s) on how to design, build, launch and maintain a production, public-facing ML MVP in Streamlit with enterprise-grade security. Prototype here (direct message me for the user name and password). Yes we signed up a $USD 400M+ turnover enterprise customer before launch. I am gauging interest in this topic 
.
Background
My startup is super excited to be launching our enterprise-grade machine learning MVP (using AWS and Render) after 6 months of development and I couldnāt be prouder to have done it using only Python and Streamlit 
. Weāve even signed up our first paying enterprise customer before launch which is 
Coming from a Python-only background, my experience with Streamlit for front-end development has been a delight and stood up to our demands 
.
Problem Motivation
I believe that as data scientists we can help our communities through building software products that users love and solve real problems. I want to motivate you to try!
But as data scientists, building enterprise-grade software can be dauntingā¦ however if I can do it, Iād like to encourage you to think you can too, if youāre willing to learn!
I am eager to contribute back to the community and keen to share our experience, lessons learned, and to learn from others.
If I can also dispel some of the persistent myths about Streamlitās inability to handle complex, enterprise grade ML products that would be great too but by no means the main objective.
Webinar
I am gauging interest from the Streamlit community in joining an interactive webinar(s) where I will be sharing our product journey in Streamlit.
Proposed topics:
- A brief background on what Iām passionate about, our company mission, our product journey so far, where we are heading
- Why doing market-user research and prototyping before product building reduces the risk of user rejection and wasted effort
- Levelling up your Streamlit game: production, architecture, security and user experience for data scientists
- Product architecture for ML apps: what it is, and why it is important
- Comparing the programming challenges for the front-end vs back-end
- Continuous Integration and Continuous Delivery: what is it and why is it useful?
- How to make your Python code base for Streamlit smaller and easier to: read, re-use and maintain
- Building in Streamlit: structuring your code, UX, achieving enterprise-grade security using OAuth 2.0 and secrets management, flow control, state management, exception handing, testing, and deployment. This could be two or more sessions depending on depth of interest.
- Maintaining and monitoring your app: MLOps and security perspectives
- Reflecting on Streamlit: why it has been PERFECT for us, our MVP use case and our startup as a business, considering other alternatives
Iām proposing that while this will be an applied session(s), I want to outline some of the design challenges involved, the tensions that arise between objectives, and some of the trade-offs you will likely weigh up as you build towards a production-grade MVP.
In true product fashion, I will likely need to iterate on the above topics and format from people. But, at this stage, I wanted to gauge potential interest before launching into it.
Timing of the first of these likely to be in early November. If there is enough interest and/or people want more depth, Iād be happy to break topics up into separate sessions.
Please like this post if you are interested and reply which topic number(s) you are most interested in. Iāll count the votes after a month or so, and structure the agenda accordingly. If thereās topics you think Iāve missed please also let me know!
- Reducing product risk with market-user research and prototyping
- Levelling up your Streamlit game
- Product architecture for ML apps
- Programming challenges for the front-end vs back-end
- Continuous Integration and Continuous Delivery
- Python code base for Streamlit
- Building in Streamlit
- Maintaining and monitoring
- Reflecting on Streamlit
Thanks!
