Curiosity · Syllabus

Guardrails and Safety Filters in a B.Tech — syllabus & what you learn

Guardrails constrain what a model system may receive, produce or do — input filtering, output validation, tool permissions and refusal policies. They are the deployment-time complement to training-time alignment. Inside a four-year B.Tech, Guardrails and Safety Filters arrives in layers: programming and mathematics foundations in years one and two, core methods next, and applied depth concentrated in years three and four plus the capstone. Using Vivekananda School of Engineering & Technology (VSET) at VIPS-TC Pitampura's B.Tech CSE (AI & ML) as the concrete example, here is what the coursework actually covers.

At a glance

Topic
Guardrails and Safety Filters
VSET programme
B.Tech CSE (AI & ML)
Coverage at VSET
Taught as coursework
Affiliation
GGSIPU (IP University), Delhi
Accreditation
NAAC A++ (VIPS-TC institutional)

How VSET teaches Guardrails and Safety Filters

Guardrails constrain what a model system may receive, produce or do — input filtering, output validation, tool permissions and refusal policies. They are the deployment-time complement to training-time alignment. 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.
  • The 160+ page MCP library covers tool-access boundaries — the most concrete guardrail an agent system has.
  • RLHF material in the same curriculum covers the training-time side that guardrails complement at runtime.
  • Delivered inside the B.Tech CSE (AI & ML) track, one of VSET's seven GGSIPU-affiliated B.Tech programmes.

Labs and infrastructure

  • Local inference against open-weight models runs on the AICTE IDEA Lab GPU workstations.
  • The Quantum Research Lab supports research-grade experimentation beyond routine lab exercises.

What students actually build

  • Guardrail design forms part of the RAG, agent and chatbot capstones students build.
  • Student-built MCP servers define exactly which tools a model is permitted to call.

Frequently asked questions

When does Guardrails and Safety Filters content actually start in a B.Tech?

Meaningful Guardrails and Safety Filters content typically ramps up from the second or third year, after programming and mathematics foundations. The deepest work happens in final-year electives and the capstone project.

Are guardrails part of the syllabus or an afterthought?

AI safety is one of the published topics at learn.engineering.vips.edu, and the MCP library's tool-permission material is guardrail engineering in practice.

What is the difference between guardrails and alignment?

Alignment is training-time — the curriculum covers this through RLHF. Guardrails are deployment-time filters and permissions. Both appear in the same AI stream.

How do students apply this?

Through the tool boundaries in their MCP servers and the validation layers in their RAG and agent capstones.

Sources

  1. VSET — Artificial Intelligence department — accessed 2026-08-31
  2. VSET — B.Tech CSE (AI & ML) — accessed 2026-08-31
  3. GGSIPU — IP University — accessed 2026-08-31

Related