Agents & Automation
Agentic Systems
Build software that plans, calls tools, and recovers from failure
An agent is not a chatbot with a longer prompt. It keeps state, calls tools, and has to recover when a step breaks. This track covers the orchestration patterns behind the agentic products people actually ship.
- How it runs
- Self-paced, applied in your capstone
- Starting point
- After AI Engineering Foundations
- Included in
- ₹9,999
03
Agents & Automation
What you actually learn
You already know how a model behaves. Here you build an agent that plans multi-step work, calls the right tool at the right time, and fails gracefully instead of silently.
Tool calling, and writing tools an agent can actually use without breaking
Memory: what to persist, what to summarise, and what to throw away
Multi-step reasoning, and graceful failure when one step goes wrong
Tracing a run so you can debug an agent instead of guessing at it
Where it lands in the 12 weeks
This is coursework you do while you are building, not before it. It is timed to hit when your capstone needs it, and it ends somewhere: a profile in front of people who are hiring.
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9
- 10
- 11
- 12
Applied through the capstone, Weeks 9 and 10
This program is part of the Launchpad
Every program here runs inside the 12 weeks LvlUp Career Launchpad, for ₹9,999. Applications for the next cohort close soon.