1 of 21

Determining your AI needs

2 of 21

Terms of Use

Except where otherwise indicated, The contents of this slide presentation are available for use under the Creative Commons Attribution 4.0 license.

You are free to adapt and share the work, but you must give appropriate credit, provide a link to the license, and indicate if changes were made.

Sample attribution: [Title of work] by Fred Hutchinson Cancer Center Data Science Lab. CC-BY 4.0

3 of 21

What do you need to use AI for?

What are your use cases?

Why do you need a custom AI tool?

CC-by hutchdatascience.org

4 of 21

Input

Output

Training Data

What is this?

It’s an apple.

Algorithms

What’s your use case?

CC-by hutchdatascience.org

5 of 21

Input

Output

Training Data

What is this?

It’s an apple.

Algorithms

What’s your use case?

Customized Knowledge

My input questions are domain specific.

CC-by hutchdatascience.org

6 of 21

Input

Output

Training Data

Can you do math on my social security number ***-**-**** ?

It’s a pear.

Algorithms

What’s your use case?

Customized Security

There is protected data involved

CC-by hutchdatascience.org

7 of 21

Input

Output

Training Data

It’s an apple.

Algorithms

What’s your use case?

Customized Interface

I need users to access an AI model but from our own website/app

What is this?

CC-by hutchdatascience.org

8 of 21

The more customization you need,

the more work that will be required by your team

Carefully consider what the project truly requires!

*But protected data needs should always be taken seriously! Before deploying any AI solutions that involve protected data consult your legal and IRB experts

Identify what is important

CC-by hutchdatascience.org

9 of 21

What do you need from this AI project?

– you may need a combination of these items!

Your solution may include:

Training an AI model

Deploying on secure server

Using an API of an existing AI platform

Needs:

Customized knowledge

Customized security

Customized interface

Note: if none of these are your needs, you may want to reconsider whether customized AI is truly needed!

https://docs.google.com/presentation/d/1GgsDe7dYA91RY_xCbcDRzSPGuxIMXnyHPbzL1h57pE4/edit#slide=id.g1ed2b5c8ee1_0_374

CC-by hutchdatascience.org

10 of 21

Other things you probably need to consider:

The technical expertise available on your team

Your funding situation

The quickness of the deadlines

CC-by hutchdatascience.org

11 of 21

Need vs. effort

Technical expertise

Customization needs

CC-by hutchdatascience.org

12 of 21

Technical expertise

Staffing

Customization	Data Handling	Back End Dev	Front End Dev	API handling	Machine learning algorithms	Secure computing
Knowledge	Required				Suggested
Security		Required	Suggested		Suggested	Required
Interface		Suggested	Required	Required

CC-by hutchdatascience.org

In this table we describe what kinds of technical expertise you will likely need on your team based on what kinds of customization eh I needs your project entails. Keep in mind you can likely minimize these staffing needs if you pay for products that are prepackaged. Prepackaged products (which we will discuss in future chapters) generally require less expertise but will not allow you the same freedom for more granular customization. For knowledge needs, you will likely require a team who is comfortable with data handling techniques. It will also be ideal if they have a certain knowledge of machine learning algorithms. For Security needs, it’s likely you will need someone comfortable with back end development and secure computing. Depending on your strategies with this need, it would also be good if you have a front end developer’s help. For customized interfaces, you’ll likely need a front end developer, likely someone who is comfortable with API handling, which also means potentially a back end developer.

13 of 21

How much would your AI tool cost per query (on average)?

How many queries might a users submit in an average session?

How big are the data the users would be inputting?

Given the answers to the above how many users would you be able to accommodate?

Evaluate pricing for the services you are using given your estimates for users and inputs.

Computing costs

CC-by hutchdatascience.org

14 of 21

Computing resource costs -

from scratch isn’t always cheaper

Graph based on: https://lajavaness.medium.com/llm-large-language-model-cost-analysis-d5022bb43e9e

Cost

Usage

AWS for custom build

API for ChatGPT

This is only computing costs! Does not take into account other costs like salaries of those maintaining the custom build!

CC-by hutchdatascience.org

15 of 21

Custom deployments will require more technical expertises on hand – think salary costs

User experience designers are highly valuable for user facing projects

Project managers will help everyone save time

Administration staffing needs should not be overlooked

Staffing costs

CC-by hutchdatascience.org

16 of 21

How quickly does this need to be ready?

What determines this?

How long does this need to be maintained?

More customized deployments will require more development time

If you rush development technical debt will be incurred

Time Assessment

CC-by hutchdatascience.org

17 of 21

Customizations	Strategy summary	Example	Technical expertise	Funding	Time
Interface	No code deploy AI features using a service	Cogniflow
Security	Privacy services for commercial AI	Private AI
Knowledge Interface	Train AI platform and deploy using an API	ChatGPT API
Knowledge �Interface Security	Train and deploy a open source model	Hugging Face

Example Project Strategies

Investment

CC-by hutchdatascience.org

Up until this point we've been discuss strategies in very vague terms. To bring this discussion to specifics, we will discuss some example AI tool project strategies you may employ based on what combination of customization AI needs you have.

These example project strategies are in the order of least to most resource and time investment.

In the leftmost column is described what kinds of customizations are able to be made given the described strategy. In the example column we have links to resources and platforms that would be a central point or product for this strategy.The technical expertise column describes vaguely how much technical expertise in house you would need to deploy the example strategy. The funding column describes approximately the funding costs that would be associated with the strategy (but not this is highly variable given specifics of a project).The time column describes how long it would take to deploy this solution.

18 of 21

Customizations	Strategy summary	Example	Technical expertise	Funding	Time
Interface	No code deploy AI features using a service	Cogniflow
Security	Privacy services for commercial AI	Private AI
Knowledge Interface	Train AI platform and deploy using an API	ChatGPT API
Knowledge �Interface Security	Train and deploy a open source model	Hugging Face

Example Project Strategies

Investment

CC-by hutchdatascience.org

19 of 21

Customizations	Strategy summary	Example	Technical expertise	Funding	Time
Interface	No code deploy AI features using a service	Cogniflow
Security	Privacy services for commercial AI	Private AI
Knowledge Interface	Train AI platform and deploy using an API	ChatGPT API
Knowledge �Interface Security	Train and deploy a open source model	Hugging Face

Example Project Strategies

Investment

CC-by hutchdatascience.org

20 of 21

Customizations	Strategy summary	Example	Technical expertise	Funding	Time
Interface	No code deploy AI features using a service	Cogniflow
Security	Privacy services for commercial AI	Private AI
Knowledge Interface	Train AI platform and deploy using an API	ChatGPT API
Knowledge �Interface Security	Train and deploy a open source model	Hugging Face

Example Project Strategies

Investment

CC-by hutchdatascience.org

21 of 21

Customizations	Strategy summary	Example	Technical expertise	Funding	Time
Interface	No code deploy AI features using a service	Cogniflow
Security	Privacy services for commercial AI	Private AI
Knowledge Interface	Train AI platform and deploy using an API	ChatGPT API
Knowledge �Interface Security	Train and deploy a open source model	Hugging Face

Example Project Strategies

Investment

CC-by hutchdatascience.org

Hugging Face is a community and repository of open source AI and machine learning models of all kinds of varieties. The resources available on Hugging Face would allow you to customize an AI tool to meet all the needs you might have. The open source nature of the AI models, datasets, and examples available on Hugging Face means that this not completely from scratch either but would require more technical expertise to utilize the resources here. The tutorials and resources on hugging face would allow you to control and build a AI model that fits every need you might have. But remember computing and staffing are always costs, hence why we have not said that this is *necessarily* a less expensive strategy. Depending on the size and technical expertise of your team it will likely take more time than the other strategies.