Agentic LLMs
Context
Modern LLMs are increasingly enhanced with the ability to interact with external tools such as:
- Code interpreters
- Search engines
- Databases
- Simulated environments
Motivation
This project investigates Agentic LLMs — models capable of reasoning, planning, and acting through tool usage in dynamic environments.
Students will work on:
- Tool-calling architectures
- Environment interaction loops
- Memory-augmented reasoning
- Autonomous problem-solving workflows
Goal
- Implement agent frameworks for LLMs
- Integrate tool APIs into inference pipelines
- Train models for multi-step decision making
- Evaluate performance on real-world tasks
Requirements
- Strong programming skills
- Interest in AI systems and agents
- Experience with Transformers and ML frameworks helpful