on day 10. Can't wait Special thanks to Ed Donner for always replying quickly to our questions.
I'm still trying to get used to this asyncio syntax ๐ฅฒ๐ฅฒ๐ฉ
#AgenticAI#LLMs
processing and giving the output
On day 9, i learnt about using the Open AI SDK inbuilt tools like searching the web.
I built a very satisfying project a researcher that gets more information on a particular query.
I will be turning this little project into an Application
I just finished day 8 and 9 in my Agentic AI course. Got my hands dirty with Guardrails. in day 8 I built an autonomous Agentic workflow where an agent asks three LLM to generate texts, it picks the best text sends the output to a handoff which is another agent to do more
Day 6 on my Agentic AI journey:
I learnt about asyncio that agents uses. Asyncio provides light weight alternative to threading and multi threading.
The main topic for the was Open AI Agents SDK .
Just did little little intro where I used the Agent to to create an instance of
In 2026, Iโm building in public.
Last year, I hid myself
This year, Iโm choosing visibility, consistency, and learning out loud.
Iโm a ML Engineer/ Data scientist sharing my work, process, and lessons as I grow
If youโre creating or pushing your career forward..lets connect.
I deployed the chat bot to hugging face spaces and will be sharing the link.
NB: The pic was when I asked my career chat bot a question.
Week 1 complete โ
5 more weeks to go ๐ช
#Agents#llmengineering#progress#dailygoals
Day 5 on my Agentic AI course
Built an Agentic Career/Personal chat bot about my self
Understood resources and tools:
These are very integral parts of AI systems as a professional said " Agent runs tools in a loop to achieve a goal"
Goals for 2026 :
โ Become strong in competitive programming
โ Deliver real, production-ready projects
โ Build a visible tech presence
โ Maintain consistency on GitHub
Hereโs to learning, building, and becoming better in 2026 ๐ซถ
prompt. If the response isn't satisfactory I had a function to rerun and give another response
It was fun to learn about resources, tools, and structured outputs. Solidifying my knowledge on validating responses from models
Let's go
#AgenticAI#LLMs#Langchain#machinelearning
external code.
At the end i built an Agentic pipeline. The first model answers questions about me based on my LinkedIn profile i provided, The second model validates the answer of the first model if the response is satisfactory after providing it my LinkedIn profile too in the
Just finished day 4 on my Agentic AI course. Learned resources and tools. Resources are just adding more information to the prompt you are giving the model. Here I downloaded my LinkedIn profile and added it to the prompt. Tools are ways we can give models options to call an
These kind of patterns to send a task to multiple LLM's and evaluate results are common where one need to improve the quality of the LLM response. This approach can be universally applied to business projects where accuracy is critical.
#agents#LLMs#learningjourney
We pushed the responses to another LLm to rank the responses from best to last.
I only used different Open AI models since i only paid for their Api and ollama
Day 3 on my Agentic AI course was a little bit hands on, The topic was LLM orchestration where we used different LLM's to solve a problem and tried to get the best response.
The tutor used gpt-4o-mini, Claude-3-7-sonnet,Gemini-2.0-flash,deep seek,Groq and Ollama