Özhan Atalı

My AI notes - 2/3

31 Eki 2024 - İstanbul

Starting an AI project

-Workflow of projects -Selecting AI projects -Organizing data and team for the projects

Workflow of a ML Project

ie speech recognition or self-driving car tools:Amazon Echo/Alexa keywords

Step1: Collect data Step2: Train model (You have to iterate many times until, hopefully, the model looks like is good enough.) Step3: Deploy model (incl. maitain/update)

Workflow of a Data Science Project

ie Optimize a sales funnel (of an onlines market) or automize a manufacturing line or ML to prioritize the leads of customers/collections or Automated visual inspections or Automated resume screening or Recommending personalized ads/services

Step1: Collect data Step2: Analyze data Step3: Suggest hypotheses/actions

**Every job function needs to learn how to use data

How to choose an AI Project

Interception of “what AI can do/AI experts” and “Valuable for your business/domain experts”

Brainstorming framework:

…much more useful to think about automating tasks rather than automating jobs… …you can make progress even without big data, even without tons of data…

1st due diligence (technical/business/ethical) the project before starting 1-Technical: -Consider AI performance -How much data needed -Engineering timeline 2-Busines: -Costs -Revenue -Launch new product/business? 3-Ethical diligence

Build or Buy??

ML can be >inhouse/outsource Data Science >inhouse if an industry standard >avoid building it

Working with an AI team

-specify an acceptance criteria for the project

of Training set > # of Stest Set

Technical tools for AI teams

Open Source: PyTorch, TensorFlow, Hugging Face PaddlePaddle, Scikit-Learn, or R Academic: Arxiv Open Repo: GitHub edge deployments:the computation has to happen usually in a computer right there inside the car, that’s called an edge deployment. edge/cloud/on-prem








📝 | Sitenin son güncellenme tarihi: 08 Kas 2024