Due to the recent changes made to the Azure AI Engineer Associate exam, I have decided to reschedule my presentation and take some time to relax. Within the past two months, I have obtained three certifications, but given the latest changes to the exam, I perceive it to be even more challenging. Furthermore, I have noticed that there is limited information available on the internet regarding the exam. To address this issue, I have reviewed the documentation and preparation page on the official site, and I highly recommend to you complete all of the lectures and labs if you're looking for your certification. In order to clarify my roadmap and decision-making when building a system with these tools, I have decided to compile a series of resources covering all of the Azure AI services.
Cognitive Services
- Vision Services
- Computer vision
- Custom vision
- Face API
- Speech Services
- Text to speech
- Speech to text
- Speech translation
- Speech recognition
- Language Services
- Entity recognition
- Sentiment analysis
- Question answering
- Conversational Language Understanding
- Translation services
- Desicion Services
- Anomaly detector
- Content moderator
- Personalizer
- OpenAI Service - Don't needed to the exam since it's a limited service.
Applied AI Services
- Form recognizer
- Video Analyzer
- Inmersive reader
- Bot service
- Cognitive search
Azure Machine Learning
- AutoML
- ML designer
Capabilities of each service
Using the rest API, you also could use the SDK available for Python and C#.
Please note that these services are subject to constant change and improvement by the Azure team, which means that this information may become outdated.
TIP: Check this page, they had a lot of questions and answers very useful for free.
Responsible AI and ethics
This is a topic that Microsoft takes seriously. You must know this 6 principles and be capable of apply each one in case studies and of course, in real life projects.
Here is my key points for each principle:
- Fairness: AI systems must treat all users equally and avoid any biases towards specific groups of people.
- Transparency: You must inform users of your AI application about its capabilities, purpose, limitations, and functioning.
- Privacy and Safety: The AI system should protect and maintain the privacy and security of personal data.
- Reliability: The system should undergo rigorous testing to ensure its reliability and accuracy.
- Inclusiveness: The system must incorporate a diverse range of data during its training to avoid biases resulting from common cases.
- Accountability: The individuals responsible for a model are accountable for its behavior.
Comments