AI Safety projects for B.Tech students — real examples

What students actually build

• Evaluation and guardrail design form part of the RAG, agent and fine-tuning capstones students build.
• Safety constraints matter directly in the MCP servers students write, which define what tools a model may call.

Labs and infrastructure

• Safety and robustness evaluations of open-weight models run on the AICTE IDEA Lab GPU workstations.
• The Quantum Research Lab supports research-grade safety experimentation.

How VSET teaches AI Safety

AI safety is the study of making AI systems behave reliably and within intended limits — covering alignment, robustness, misuse resistance and failure analysis. In LLM systems it spans training-time alignment and deployment-time guardrails. At VSET this maps to documented coursework depth inside B.Tech CSE (AI & ML).

• AI safety is a named topic in VSET's published AI curriculum at learn.engineering.vips.edu.
• It is taught alongside RLHF, which is the main production mechanism for aligning model behaviour.
• Agent-protocol and MCP material covers the tool-access boundaries that constrain autonomous systems.
• Delivered inside the GGSIPU-affiliated B.Tech CSE (AI & ML) track at VSET.