This repository accompanies the Agentic AI course designed and delivered by Anupam Purwar.
The course focuses on building job-ready, industry-relevant skills in modern AI systems—moving beyond chatbots to autonomous, tool-using, decision-making AI agents.
The curriculum is aligned with real industry needs, startup ecosystems, and national priorities around AI talent and innovation.
Traditional AI education often stops at models and theory.
Agentic AI goes further:
- LLMs that plan, reason, and act
- AI systems that use tools, APIs, memory, and external knowledge
- Architectures used in real production systems (not demos)
- Skills directly applicable to jobs, startups, and research
This course treats LLMs as a computational core, not just text generators.
Mental model
- LLM = CPU
- Prompt + context = program
- Tools = I/O devices
- Memory = working & long-term state
- Agent loop = operating system for intelligence
- Conversational AI vs Agentic AI
- Foundations of AI agents and levels of agency
- Natural Language Understanding (NLU)
- Transformers and Large Language Models (LLMs)
- Prompt engineering for reasoning and control
- Retrieval Augmented Generation (RAG)
- Tool-calling, planning, memory, and autonomy
- Multi-agent systems and orchestration
- Evaluating agents and mitigating risks
- Designing agent architectures
- Building real AI agents (not toy chatbots)
- Using agents for:
- Recruitment & HR
- Knowledge assistants
- Decision support systems
- Automation workflows
- Understanding cost, ROI, and deployment trade-offs
- Foundations of Conversational & Agentic AI
- Advanced NLU for Agents
- LLMs as the Core of Agentic Systems
- Agent Architectures & Components
- Prompt Engineering & RAG
- Building & Implementing AI Agents
- Advanced Agentic Concepts & Multi-Agent Systems
- Business Considerations & Future of Agentic AI
This course is explicitly designed to improve employability:
- Focus on how companies actually build AI systems
- Skills mapped to:
- AI Engineer
- Applied ML Engineer
- Agent / Automation Engineer
- AI Product & Solutions roles
- Strong emphasis on:
- System thinking
- Trade-offs and design decisions
- Real-world constraints (cost, latency, reliability)
Students finish the course able to explain, design, and build agentic systems—not just use libraries blindly.
This course is conceptually aligned with the POET Framework, a strategic blueprint for building a sustainable and sovereign AI ecosystem in India.
POET stands for:
- P – Protect our Indic Data
- O – Opportunities
- E – Employable Youth
- T – Training Infrastructure
The framework emphasizes that AI progress is meaningless without skilled people who can build and deploy systems.
- Teaches responsible AI system design
- Emphasizes evaluation, risk mitigation, and safe deployment
- Encourages indigenous capability building, not dependency on black-box tools
- Agentic AI enables:
- Startups
- Applied research
- Domain-specific AI solutions
- Students learn how to convert AI capability into real economic value
- Bridges the gap between academia and industry
- Focuses on skills companies actually hire for
- Prepares learners for a future where:
- Pure prompt usage is not enough
- System builders outperform tool users
- Introduces modern AI stacks and workflows
- Prepares learners to work with:
- Cloud AI systems
- Scalable agent architectures
- Tool-based AI pipelines
Anupam Purwar
AI Engineer, Educator & Researcher
- Focus areas: Agentic AI, LLM Systems, Applied AI
- Experience bridging academia, industry, and startups
- Strong emphasis on clarity, first principles, and real-world relevance
🔗 Profile: https://anupam-purwar.github.io/page/
- Students with basic Python & ML knowledge
- Engineers transitioning into AI roles
- Researchers exploring applied agent systems
- Professionals looking to future-proof their skills
- Educators designing modern AI curricula
This repository may include:
- Lecture notes & slides
- Diagrams & explanations
- Example agent architectures
- Sample projects & demos
- References and further reading
It is intended as a learning companion, not just code dumps.
Build AI systems that think, act, and deliver value—
while creating a generation of employable, system-level AI engineers.
Aligned with the spirit of Atmanirbhar AI and a future-ready workforce.
- Open-source AI community
- Research from industry & academia
- Learners and practitioners pushing AI beyond chatbots
If you find this useful, consider ⭐ starring the repository and sharing it with learners who want to build real AI systems.