Introduction: This 5 day introduction course is very helpful for the seekers who want to go ahead in the world of A.I. Personally what i believe is this course is very well made for those people who have the some background in any of the computer language and Deep learning and want to know about Agents. In this course you will find the question of these answers: 1: What are the agents
2: why we need the agents
3: Are these really necessary for the current real life scenario
4: How does it works
5: learn to make your agent
Here is the brief summary of what we have done so far:
Day 1
Introduction to Agents: learned how we import the tool kit to create the our agent. We can consider our agent as our robot which we have to assemble and all the ingredients comes in package now our work is to assemble it from that package and build a robot which is not physical but the abstract one.
so in brief we learn how to assemble our robot(agent) which specializes in the particular stuff according to the user what he build it for and uses the google search as its tool.
Day 2
Agent Tools & Interoperability with Model Context Protocol: previously we used the google search as a tool for our agent so in day we learned how we make our custom tools according to our need. learned to connect with A.I using function tool so it will look for the data which is not in the training set, learning to make our agent more reliable such that A.I. stop to hallucinate to stop guessing and do actual maths doing actual coding. And finally learned how to make agent as tool such that this particular agent will used as a custom tool for the larger agent.
Day 3
Agent Sessions: If we talk about the raw LLM, in core these are stateless such that if you ask for a solution it will only provide for the current problem's solution without considering the past conversation. So in this particular day we learned how we can manage the session using agents. we learned how we can provide the short term and long term memory management using agent and how can we allow our agents to do so.
Day 4
Observability: learned how it will allow us to see how our agent is taking decision such that we can find out the exact reason and rectify it in short ways to identify and debug. when our agent is in developing phase how debug and after development how can we use pulgin to debug. And also learned about the metrics which allows to check how good our agent is performing
Day 5
Agent to agent communication: A single agent cant do much, it have its own limitation so to overcome this issue we learned about how to establish the communication between multiple agents who specializes in different domain such that it combines all and give the output as user's need .
Top comments (0)