From LLMs to Agentic AI | Hands-on workshop
From LLMs to Agentic AI is a 3-day, in-person, hands-on workshop on building real-world agentic systems with LLMs.
The workshop is deliberately built around a single argument: an LLM is not an agent, and most "agent" products in the wild are confused about which one they are. Over three days we replace that confusion with a working mental model and the code to back it up. Participants progress from LLMs as conditional token generators, through retrieval as the first control primitive, through reasoning models, into agents — built three different ways and compared — and finally into multi-agent systems with MCP-based tool integration.
It is highly interactive and technical. Short lectures build the mental model; tutorials connect ideas to concrete implementations; "you do it" labs expose the failure modes and trade-offs that only show up in code. The goal is not exposure but understanding — and the engineering judgment to know when an agent is the right answer at all.
Language of instruction: English.
Requirements
Participants should be comfortable coding in Python and have a practical understanding of LLMs; no prior agent-framework experience is required.
This workshop is for participants who:
Are working engineers, ML engineers, data scientists, technical leads, or applied researchers.
Are proficient in Python and comfortable on the command line.
Know what an LLM is at a prompting level and have used APIs such as OpenAI, Anthropic, or Hugging Face.
Do not need prior agent-framework experience — we start from first principles.
Have basic comfort with technical abstractions and algorithmic thinking.
Dates
Thursday, Oct 22: 9:00 AM - 6:00 PM
Friday, Oct 23: 9:00 AM - 6:00 PM
Saturday, Oct 24 8: 9:00 AM - 1:00 PM
Location: Universidad de Concepción, Campus Santiago (Marchant Pereira 10, Providencia, Santiago)
Main Instructor
Dr. Pavlos Protopapas: Scientific Program Director, Harvard John A. Paulson School of Engineering and Applied Sciences.
Speakers
Guillermo Cabrera: Director of CDIA - UdeC.
Shivas Jayaram: Co-Founder & CTO, INSTEL.
Ignacio Becker: Senior ML Engineer at CMPC.
Karim Pichara: Co-founder of NotCo.
Price
Standard Rate: $1,000,000 CLP
25% early bird discount: available until August 28.
20% UdeC Alumni discount.
20% Institutional discount: for organizations registering 2 or more participants.
What is covered:
Full access to sessions, labs, and materials
Catering & lunch
Certificate of completion
Ongoing community access
Payment Methods
Bank transfer
Credit card
Debit card
Registration process
Register for the event.
Pay the registration fee at the following link: https://bit.ly/4gsuEQD
Send proof of payment to cdia@inf.udec.cl to aprove your registration.
For more information, contact us via Whatsapp in this link.